high temperature materials characterization and sensor
TRANSCRIPT
University of Central Florida University of Central Florida
STARS STARS
Electronic Theses and Dissertations, 2004-2019
2012
High Temperature Materials Characterization And Sensor High Temperature Materials Characterization And Sensor
Application Application
Xinhua Ren University of Central Florida
Part of the Electrical and Electronics Commons
Find similar works at: https://stars.library.ucf.edu/etd
University of Central Florida Libraries http://library.ucf.edu
This Doctoral Dissertation (Open Access) is brought to you for free and open access by STARS. It has been accepted
for inclusion in Electronic Theses and Dissertations, 2004-2019 by an authorized administrator of STARS. For more
information, please contact [email protected].
STARS Citation STARS Citation Ren, Xinhua, "High Temperature Materials Characterization And Sensor Application" (2012). Electronic Theses and Dissertations, 2004-2019. 2304. https://stars.library.ucf.edu/etd/2304
HIGH TEMPERATURE MATERIALS CHARACTERIZATION AND SENSOR
APPLICATION
by
XINHUA REN
B.S. University of Science and Technology of China, 2006
M.S. University of Central Florida, 2010
A dissertation submitted in partial fulfillment of the requirements
for the degree of Doctor of Philosophy
in the Department of Electrical Engineering and Computer Science
in the College of Engineering & Computer Science
at the University of Central Florida
Orlando, Florida
Fall Term
2012
Major Professor: Xun Gong
ii
© 2012 Xinhua Ren
iii
ABSTRACT
This dissertation presents new solutions for turbine engines in need of wireless temperature
sensors at temperatures up to 1300oC. Two important goals have been achieved in this
dissertation. First, a novel method for precisely characterizing the dielectric properties of high
temperature ceramic materials at high temperatures is presented for microwave frequencies. This
technique is based on a high-quality (Q)-factor dielectrically-loaded cavity resonator, which
allows for accurate characterization of both dielectric constant and loss tangent of the material.
The dielectric properties of Silicon Carbonitride (SiCN) and Silicoboron Carbonitride (SiBCN)
ceramics, developed at UCF Advanced Materials Processing and Analysis Center (AMPC) are
characterized from 25 to 1300oC. It is observed that the dielectric constant and loss tangent of
SiCN and SiBCN materials increase monotonously with temperature. This temperature
dependency provides the valuable basis for development of wireless passive temperature sensors
for high-temperature applications.
Second, wireless temperature sensors are designed based on the aforementioned high-
temperature ceramic materials. The dielectric constant of high-temperature ceramics increases
monotonically with temperature and as a result changes the resonant frequency of the resonator.
Therefore, the temperature can be extracted by measuring the change of the resonant frequency
of the resonator. In order for the resonator to operate wirelessly, antennas need to be included in
the design. Three different types of sensors, corresponding to different antenna configurations,
are designed and the prototypes are fabricated and tested. All of the sensors successfully perform
at temperatures over 1000oC. These wireless passive sensor designs will significantly benefit
turbine engines in need of sensors operating at harsh environments.
iv
ACKNOWLEDGMENTS
I wish to express appreciation to my advisor, Prof. Xun Gong, for his guidance and support over
my six years at the University of Central Florida. He provided me with a project that turned out
to become the basis of my dissertation He taught me many things both inside and outside of the
classroom, and I will get benefits from these lessons forever. The members of my dissertation
committee: Prof. Thomas Wu, Prof. Parveen Wahid, and Prof. Linan An for sharing their wealth
of knowledge through courses, valuable insight and suggestions on how to improve the quality of
my research. Thank you!
The members of the Antenna, RF, Microwave, and Integrated Systems (ARMI) lab for the
abundance of peer support. I would like to acknowledge Dr. Siamak Ebadi, Mr. Haitao Cheng ,
Mr. Justin Luther, Ms. Ya Shen, Mr. Kalyan Karnati, and Ms. Tianjiao Li for their always
valuable insights and contributions, and more importantly for their personal friendships. Thank
you!
Lastly, I wish to thank my family and friends, who have been incredibly supportive of me.
Without all of you, I certainly would not have made it this far. Thank you!
v
TABLE OF CONTENTS
LIST OF FIGURES ..................................................................................................................... viii
LIST OF TABLES ....................................................................................................................... xvi
CHAPTER 1 INTRODUCTION ................................................................................................. 1
1.1 Motivation ........................................................................................................................ 1
1.2 Literature Review: Existing Material Characterization Methods .................................... 4
1.3 Overview of Dissertation ............................................................................................... 10
CHAPTER 2 HIGH TEMPERATURE MATERIALS ............................................................. 11
2.1 SiCN Ceramic Material .................................................................................................. 11
2.2 Fabrication Procedure .................................................................................................... 11
2.3 Dielectric Property of SiCN at DC and Low Frequencies ............................................. 13
2.4 Conclusion ...................................................................................................................... 14
CHAPTER 3 ROOM TEMPERATURE CHARACTERIZATION ......................................... 15
3.1 Introduction .................................................................................................................... 15
3.2 X-band Adapter Cavity Characterization ....................................................................... 16
3.2.1 Characterization Setup ............................................................................................ 16
3.2.2 Measurement Results .............................................................................................. 21
3.3 Ku-band Adapter Cavity Characterization ..................................................................... 22
3.3.1 Characterization Setup ............................................................................................ 24
vi
3.3.2 Measurement Results .............................................................................................. 26
3.4 Conclusion ...................................................................................................................... 28
CHAPTER 4 HIGH TEMPERTURE CHARACTERIZATION .............................................. 29
4.1 Introduction .................................................................................................................... 29
4.2 Proposed High-Temperature Characterization Method ................................................. 30
4.3 Characterization Up to 500oC ........................................................................................ 35
4.3.1 Measurement Setup Design .................................................................................... 35
4.3.2 TRL Calibration Kit ................................................................................................ 38
4.3.3 Metallization ........................................................................................................... 40
4.3.4 High Temperature Measurement Setup .................................................................. 41
4.3.5 Measurements Results ............................................................................................ 43
4.4 Characterization Up to 1000oC ...................................................................................... 48
4.4.1 Metallization Process and Measurement Setup ...................................................... 51
4.4.2 Measurement Results .............................................................................................. 52
4.5 Characterization Up to 1300oC ...................................................................................... 56
4.6 Conclusion ...................................................................................................................... 61
CHAPTER 5 HIGH TEMPERATURE MATERIALS APPLICATIONS ............................... 62
5.1 Introduction .................................................................................................................... 62
5.2 Two-antenna Sensor ....................................................................................................... 62
vii
5.2.1 Wireless Sensing System Using Passive Microwave Resonators........................... 62
5.2.2 Sensor Design ......................................................................................................... 68
5.2.3 Fabrication and Measurement up to 500oC ............................................................. 73
5.2.4 Fabrication and Measurement up to 1000oC ........................................................... 78
5.3 One-antenna Sensor........................................................................................................ 81
5.3.1 Sensor Design ......................................................................................................... 81
5.3.2 Fabrication and Measurement up to 1000oC ........................................................... 84
5.4 Integrated Slot Antenna Sensor ...................................................................................... 89
5.4.1 Sensor Design ......................................................................................................... 89
5.4.2 Fabrication and Measurement up to 1000oC ........................................................... 97
5.4.3 Fabrication and Measurement up to 1300oC ......................................................... 103
5.5 Conclusion .................................................................................................................... 111
CHAPTER 6 SUMMARY AND FUTURE WORK ............................................................... 112
6.1 Summary ...................................................................................................................... 112
6.2 Future Work ................................................................................................................. 113
LIST OF REFERENCES ............................................................................................................ 114
viii
LIST OF FIGURES
Figure 1.1: Illustration of the need for sensors to monitor extremely high temperatures of the
turbine blade; novel material characterization methods and wireless sensing mechanisms are
necessary to achieve this goal. ........................................................................................................ 2
Figure 1.2: Different transmission-based material characterization methods which can be used
for high temperature material characterization; (a) free space, (b) capacitive, (c) rectangular /
circular waveguide, and (d) coaxial line. ........................................................................................ 5
Figure 1.3: Different reflection based material characterization methods which can be used for
high temperature material characterization; (a) rectangular/circular waveguide, and (b) coaxial
probe. .............................................................................................................................................. 7
Figure 1.4: Different resonator based material characterization methods which can be used for
high temperature material characterization; (a) rectangular sample with the same cross-section
dimension, and (b) cylindrical sample with arbitrary dimensions. ................................................. 9
Figure 2.1:Process flow of SiCN ceramics in schematics ............................................................ 12
Figure 2.2: Process flow of SiCN ceramics in photos. ................................................................. 12
Figure 2.3:Scanning electron microscope (SEM) images of the SiCN material at 500× zoom level.
....................................................................................................................................................... 13
Figure 2.4: Temperature dependence of the dielectric constant of sample SiCN (sintered at
1200oC) measured at frequency of 1, 10, and 100 KHz over temperature range of 20 ~ 675 ºC
[32]. ............................................................................................................................................... 14
Figure 3.1: Measurement setup for SiCN sample. ........................................................................ 17
Figure 3.2: Measurement result for SiCN sample. ....................................................................... 18
ix
Figure 3.3: E field of the first five resonant modes for the SiCN sample ..................................... 20
Figure 3.4: Illustration of the waveguide cavity method. For 5880 measurement, Ds=14; Hs= 3.19;
Di_r=8; Do_r=9; Hr=3.19. For SiCN measurement, Ds=9.29 Hs= 6.26; Di_r=6; Do_r=7.08; Hr=0.08
(unit: mm). .................................................................................................................................... 24
Figure 3.5: Waveguide cavity method setup. View of (a) the half cavity at the assembly plane
and (b) the whole cavity after assembly. ...................................................................................... 25
Figure 3.6: S21 results of 10 measurements for SiCN sample characterization. ........................... 27
Figure 4.1: Flowchart of the high temperature characterization mechanism. ............................... 29
Figure 4.2: Illustration of the high temperature characterization method .................................... 31
Figure 4.3: The proposed high temperature material characterization setup. ............................... 32
Figure 4.4: Details of the proposed material characterization mechanism including a CPW line
loaded with a cavity resonator. ..................................................................................................... 34
Figure 4.5:: Dimensions of (a) CPW line. (b) fabricated SiCN cavity resonator and (c) connector.
S = 1.5; G = 0.4; W = 10; T = 0.635; d = 70; D = 9.41; H = 5.01; H1 = 1; W1 = 3; L1 = 9.5; L2 =
7.92; D1 = 3.9; D2 = 0.76; S1 = 0.81. (unit: mm). ......................................................................... 36
Figure 4.6: Effect of the coupling slot dimensions on (a) resonant frequency, (b) unloaded Q, and
(c) transmission level. ................................................................................................................... 37
Figure 4.7:Designed TRL calibration kit at 12.4 GHz. (a) Through, (b) Reflect, and (c) Line. d =
70, L = 2.57. (Dimensions in mm) ................................................................................................ 39
Figure 4.8: Fabricated TRL calibration standards. ....................................................................... 39
Figure 4.9: Voltage and current density variations during gold electroplating. ........................... 41
x
Figure 4.10: (a) The connector before soldering. (b) The furnace containing high-temperature
measurement setup connected to the RF cables. (c) Sample over CPW line outside the furnace. (d)
Sample and CPW line inside the furnace. (e) Sample inside the furnace with thermal isolation at
the open ends of the furnace. ........................................................................................................ 42
Figure 4.11: Measured S21 versus frequency at different temperatures. ....................................... 44
Figure 4.12: Measured fr and Qu versus temperature. ................................................................... 45
Figure 4.13: Extracted εr and tanδ versus temperature. ................................................................ 45
Figure 4.14: Measured and simulated S21 results of the SiCN sample characterization at 25 and
500oC............................................................................................................................................. 46
Figure 4.15:Measured fr versus temperature at the center temperature of (a) 100oC, (b) 200
oC, (c)
300oC, (d) 400
oC, and (e) 500
oC. .................................................................................................. 47
Figure 4.16:Dimensions of (a) the Si4B1CN (1000oC) resonator and the coupling slot, (b) high
temperature connector and (c) holder. D = 9.37; H = 4.71; W1 = 3; H1 = 1; W = 1.5; L = 1.35; S =
0.4; L1 = 12.7; L2 = 8.64; D1 = 3.05; D2 = 1.32; D3 = 2.54; L3 = 15.24; L4 = 8.64; L5 = 4.57; L6 =
7.00; L7 = 10.92; H2 = 5.33; H3 = 8.64; H4 = 2.60; H5 = 1.00; H6 = 6.31. (unit: mm) ................. 49
Figure 4.17: (a) The furnace containing high-temperature measurement setup connected to the
high temperature cables. (b) Sample over CPW line outside the furnace. (c) Connector with
holder before soldering .(d) Sample and CPW line inside the furnace. (e) High temperature
connectors soldered onto the CPW line with the high temperature solder. .................................. 50
Figure 4.18:TRL calibration standards for the measurement ....................................................... 51
Figure 4.19:Measured S21 versus frequency at different temperatures. ........................................ 53
Figure 4.20:Measured fr and Qu versus temperature. ................................................................... 53
xi
Figure 4.21: Extracted εr and tanδ versus temperature. ................................................................ 54
Figure 4.22: Measured and simulated S21 results of the Si4B1CN sample characterization at
1000oC........................................................................................................................................... 54
Figure 4.23: (a) Schematic of the resonator over CPW line (b) dimensions of the coupling slots
(c) photos of sidewall and bottom with coupling slots for sample Si4B1CN (1200oC). (D=9.37;
H=4.71; W1=3; H1=1; W=1.5; L=1.35; S=0.4) (unit: mm) ............................................................ 57
Figure 4.24:S21 for temperature from 25oC to 1300
oC. ................................................................ 58
Figure 4.25:(a) Extracted fr and (b) extracted Qun versus temperature. ........................................ 59
Figure 4.26:(a) Extracted εr and (b) extracted tanδ versus temperature . ..................................... 60
Figure 5.1:(a) Illustration of a wireless passive sensor system. (b) Equivalent circuit of the sensor
system in (a). ................................................................................................................................. 63
Figure 5.2: Wideband antenna geometry. (G=48mm, L=28mm,W=24.2mm,S=0.15mm,h=14mm,
g=0.2mm, Wf=3mm) ..................................................................................................................... 64
Figure 5.3:4.02 GHz resonator geometry. (W1=5.3mm, W2=3mm, W3=0.3mm, W4=0.5mm,
g1=0.5mm, g2=1.6mm, g3=0.2mm, g4=0.15mm, g5=1.95mm) ..................................................... 64
Figure 5.4:Measured S21 for the 3.59 GHz resonator only and the wireless system with antenna
spacing of 3 cm, 6cm, 8 cm. ......................................................................................................... 66
Figure 5.5:Measured S21 for the 4.02 GHz resonator only and the wireless system with antenna
spacing of 3 cm, 6cm, 8 cm. ......................................................................................................... 66
Figure 5.6:Direct transmission between Antennas 1 and 2 for Co-Pol and X-Pol. ...................... 67
Figure 5.7:Comparison of wireless sensing between Co-Pol and X-Pol cases with an antenna
spacing of 4 cm. ............................................................................................................................ 67
xii
Figure 5.8:Flowchart showing the wireless sensing mechanism. ................................................. 69
Figure 5.9:(a) Passive temperature sensor with a SiCN cavity resonator coupled to CPW lines. (b)
Top and (c) side view of the sensor. S = 1.5; G = 0.4; W = 10; T = 0.635; L = 149.4; D = 9.41; H
= 5.01; H1 = 1; W1 = 3 (unit: mm) (d) Patch antenna. Wa = 9.2; La = 7.7; Lm = 2.6; Ws = 0.4; Wf=
2.45.(unit: mm) ............................................................................................................................. 70
Figure 5.10:Simulated S11 of the patch antenan shown in Fig. 5.9(d) and S21 of the structure
shown in Fig. 5.9(a). ..................................................................................................................... 71
Figure 5.11:Entire wireless sensing system. ................................................................................. 72
Figure 5.12:Simulation result for the sensing system at the distance of 20mm between antennas
in Fig. 5.11. ................................................................................................................................... 73
Figure 5.13:(a) Fabricated cavity resonator over the CPW line (b) fabricated patch antenna; all
four antennas are the same ............................................................................................................ 74
Figure 5.14:Measured and simulated S11 of the patch antenna shown in Fig. 9(d) and S21 of the
structure shown in Fig. 9(a). ......................................................................................................... 75
Figure 5.15: Final wireless passive sensing setup. ....................................................................... 76
Figure 5.16: Measured S21 with or without the antennas. ............................................................. 76
Figure 5.17: Measured S21 for sensing distances from 10 to 45 mm. ........................................... 77
Figure 5.18: Measured S21 for different temperatures. ................................................................. 77
Figure 5.19: (a) Passive temperature sensor with a Si4B1CN (1000oC) cavity resonator coupled to
CPW lines, (b) side view of the sensor, and (c) wireless passive sensing setup for room
temperature. T = 0.635; D = 9.37; H = 4.71; H1 = 1; W1 = 3 (unit: mm). ..................................... 79
xiii
Figure 5.20: (a) Measured S21 for different temperatures, extracted (b) fr and (c) Qun versus
temperature. .................................................................................................................................. 80
Figure 5.21: Schematic of the wireless temperature sensor. D=9.2; H=4.7; W1=3; H1=1 (unit:
mm). .............................................................................................................................................. 81
Figure 5.22: Simulated S11 of the patch antenna for successive distances away from the sensor
after TD gating. ............................................................................................................................. 83
Figure 5.23: Simulated resonant frequency decreases with the increase of dielectric constant of
the SiBCN material. ...................................................................................................................... 84
Figure 5.24:Dimensions of the patch antenna. (a) Top view, (b) Side view, and (c) Fabricated
antenna. L=8.3;W=10.8;D1=1.27;D2=4.32;D3=6.5;T1=1.57 (unit: mm). .................................. 85
Figure 5.25: (a) Sensor outside the furnace, (b) Measurement setup with furnace, and (c)
Resonator and CPW line inside the furnace. ................................................................................ 87
Figure 5.26:Measured S11 responses of the sensor at different temperatures Si4B1CN (1000oC). 88
Figure 5.27: Measured resonant (a) frequency and (b) Q of the Si4B1CN (1000oC) sensor ......... 88
Figure 5.28:Schematic of the wireless temperature sensor. .......................................................... 89
Figure 5.29:Circuit model for the integrated slot antenna/resonator. ........................................... 91
Figure 5.30:Top view (a) and 3-D view (b) of the integrated slot antenna and cavity resonator.
The two coaxial ports of weak coupling are for the slot antenna design. w = 0.45 mm, r = 0.174
mm, R = 0.4 mm and L = 2 mm. ................................................................................................... 92
Figure 5.31:Simulated external Q factor for different slot antenna positions d and length La. .... 93
Figure 5.32:Simulated S11 of the OEWG for successive distances away from the sensor before
TD gating. ..................................................................................................................................... 94
xiv
Figure 5.33:Simulated TD responses of OEWG with and without the sensor for 20 mm distance.
....................................................................................................................................................... 94
Figure 5.34:Simulated S11 of the OEWG for successive distances away from the sensor after TD
gating............................................................................................................................................. 95
Figure 5.35:Simulated resonant frequency decreases with the increase in dielectric constant of
the SiCN material. ......................................................................................................................... 96
Figure 5.36:Simulated resonant frequency decreases with the increase in dielectric constant of
the SiCN material for 30 mm distance .......................................................................................... 97
Figure 5.37:Slot antenna fabrication by milling ........................................................................... 98
Figure 5.38: (a)Top and (b) side views of the sensor and dimensions. (Si6B1CN (1000oC):
D=8.59, H=4.48, d=2.0, w=0.45, L=6.8; Si4B1CN (1100oC): D=9.24, H=4.82, d=2.0, w=0.45,
L=7.0) (unit: mm) ......................................................................................................................... 98
Figure 5.39: (a) Measurement set-up (b) The sensor is placed inside the heat pad with alumina
board cover. (c) Inside of the heater without alumina board cover. ........................................... 100
Figure 5.40:Measured S11 responses of the OEWG at different temperatures for (a) Si6B1CN
(1000oC) and Si4B1CN (1100
oC). ............................................................................................... 102
Figure 5.41:Measured resonant frequency and Q of the Si6B1CN (1000oC) sensor ................... 103
Figure 5.42:Measured resonant frequency and Q of the Si4B1CN (1100oC) sensor ................... 103
Figure 5.43:Schematic of the measurement setup for the slot antenna sensor ........................... 104
Figure 5.44:(a) Dimension and (b) simulated S11 result of the antenna. G=23; L1=8.4; L2=8;
W=6.4; h=1.9; s= 0.6; Wf=2.5. (Unit: mm) ............................................................................... 105
xv
Figure 5.45:(a) Fabricated PCB antenna and the high temperature sensor and (b) measurement
result. ........................................................................................................................................... 106
Figure 5.46: (a) Fabricated high temperature antenna and (b) S11 response of the antenna ....... 107
Figure 5.47:Si4B1CN (1100oC) sensor and high temperature antenna (a) outside the furnace and
(b) inside the furnace. ................................................................................................................. 108
Figure 5.48:Measured S11 responses of the Si4B1CN (1100oC) sensor at different temperatures.
..................................................................................................................................................... 109
Figure 5.49:Comparison between measured S11 responses of the Si4B1CN (1100oC) sensor at
1300oC and 1350
oC. .................................................................................................................... 109
Figure 5.50:Measured resonant (a) frequency and (b) Q of the Si4B1CN (1100oC) sensor ........ 110
xvi
LIST OF TABLES
Table 3.1: Vf of different modes for SiCN sample ........................................................................ 21
Table 3.2: Summary of results for SiCN ...................................................................................... 22
Table 3.3: Average and standard deviation (ζ) of εr and tanδ ...................................................... 22
Table 3.4: Measured fr_w, QU_W and extracted εr, tanδ with standard deviations (ζ) and
uncertainty (su) using waveguide cavity method at room temperature ......................................... 27
Table 4.1: Measured fr, Qu and extracted εr, tanδ with standard deviations (ζ) and uncertainty (su)
for SiCN sample at different temperatures ................................................................................... 48
Table 4.2: Measured fr, Qu and extracted εr, tanδ with standard deviations (ζ) and uncertainty (su)
for Si4B1CN (1000oC) sample at different temperatures .............................................................. 55
1
CHAPTER 1 INTRODUCTION
1.1 Motivation
New generations of turbines require the rotating blades to survive temperatures up to 1300oC and
more. This extremely high temperature is required in different types of turbines, including gas
and jet turbines, in order to achieve higher engine efficiency [1]. This elevated temperature
imposes a great challenge on the fabrication of turbine blades. In addition, if the temperature is
not properly monitored at the vicinity of the turbine blades, there is a high chance for
overheating, causing thermal damage. This has caused reduced engine lifetimes and calls for
more frequent replacement of the blades. “However, currently there is no existing technology
that can provide online, real time monitoring of the temperature within the hot sections of
turbines. In the past decade, several techniques were under development for this purpose” [2].
Optical-based non-contact technology is currently being developed for measuring the operating
temperature in turbine engines. This indirect measurement technology however lacks the
necessary capabilities or accuracy [3][4][5][6] and “typically fails over the time due to the
smearing of the dielectric windows” [2] . Another effective alternate method to measure these
parameters without disturbing the working environment could be to use miniaturized sensors.
Silicon carbide (SiC) and silicon nitride (Si3N4)-based ceramic micro-sensors are currently under
the development for high temperature sensing application in harsh environment
[7][8][9][10][11][12]. Those sensors are seriously restricted by limited fabrication methods, high
cost, and limited operating temperature range (typically < 800oC). In fact, any micro-sensor that
involves electronic components and packaging has a slim chance of surviving in high-
2
temperature environments. Therefore, it is highly desired to develop low profile temperature
sensors seamlessly integrated with the turbine blades, which continuously monitor the
temperature, and wirelessly communicate with the outside environment. The next-generation
sensors for turbines must have the following advantages: 1) wireless; 2) high accuracy; 3)
robustness; 4) small size and flexibility; 5) low cost. “These requirements call for new sensor
materials and innovative sensor designs” [2] .
Figure 1.1: Illustration of the need for sensors to monitor extremely high temperatures of the
turbine blade; novel material characterization methods and wireless sensing mechanisms are
necessary to achieve this goal.
Intermediate Steps:
Expected result:
High-T wireless
sensors
1hr@250oC
4hr@350 oC
ball milling
4hr@1400oC
co
mp
ress
Precursor
In N2
SiCN1- Novel materials properly
characterized at high-T
Need: High-T Sensors enabling real-time
monitoring of turbine blades
Resonator
Antenna
Antenna
Temperature
variation
Temperature
is sensed
2- Novel wireless sensing methods
3
Fig. 1.1 illustrates the problem statement and the proposed solution. Hot and worn blades are
visible in the left hand side image. This undesired fact can be prevented by developing the
aforementioned sensor concept. Such a sensor is in need of two technologies in order to be
successfully realized. First, novel materials need to be developed that can withstand the
extremely hot environment inside turbines. In addition, such materials need to have
electromagnetic properties which are desired in the sensor development. This may include
temperature-variant dielectric permittivity, which is fundamentally congruent with wireless
sensor realization. The critical final step in the material development is the electromagnetic
characterization at elevated temperatures. The results of such characterization would reveal
usefulness of the material for the sensor application.
The second question to be asked is how to use the newly developed and characterized materials
and design a sensor. This calls for development of novel wireless sensing mechanisms which are
tailored to our specific turbine application. In order for the sensor to be wirelessly interrogated, it
will need an antenna which can send out and receive electromagnetic fields containing
information about the sensor’s temperature.
Successful answers to the aforementioned two questions will result in a wireless temperature
sensor which can provide real-time monitoring of the turbine blade at extremely high
temperatures. This dissertation is initially focused on the newly proposed characterization
method for measuring the dielectric property of the new ceramic materials at high temperature.
Furthermore, several types of wireless high temperature sensors are designed and tested utilizing
4
the characterized ceramic materials. In the following section, different material characterization
techniques which can be potentially used at elevated temperatures are reviewed.
1.2 Literature Review: Existing Material Characterization Methods
Microwave frequencies are preferred for efficient transmission of electromagnetic waves, which
is critical for realization of a wireless sensor. In addition, operating at high frequencies in the
microwave regime allows for implementation of small size sensors with minimum undesired
effects on the aerodynamic performance of the turbine blades. In order to design a wireless
sensor at a particular frequency, the fundamental properties of the material, referred to as a
dielectric in this context, need to be known. The two main properties to be characterized through
this process are the dielectric constant εr and the loss tangent tanδ.
There exist several of methods which can measure the properties of an unknown material at
microwave frequencies. In all of these approaches, the unknown material is exposed to a known
electromagnetic waveform and modifies the properties of the waveform. The changes caused by
the presence of the unknown material are measured and then translated to the desired dielectric
properties using post-processing techniques. This process needs to be carried out over the
temperature range in which the turbine blade is operating.
5
Figure 1.2: Different transmission-based material characterization methods which can be used
for high temperature material characterization; (a) free space, (b) capacitive, (c) rectangular /
circular waveguide, and (d) coaxial line.
The first category of the characterization methods shown in Fig. 1.2 is based on transmission of
an electromagnetic wave between two terminals. The unknown material will be placed in the
Hot zone
Transmit
AntennaReceive
Antenna
Sample
Under TestFree
Space
Free
Space
Network Analyzer
Hot
zone
Sample
Under Test
Meter
(a) (b)
(c) (d)
Hot
zone
Sample
Under Test
Network Analyzer
Rectangular or
Circular Waveguide
Hot
zone
Sample
Under Test
Network Analyzer
Coaxial
Waveguide
6
middle of this channel and will affect its transmission coefficient accordingly. The simplest
demonstration of this concept can be made using free space transmission as shown in Fig. 1.2(a).
The electromagnetic wave is sent out and received by two antennas which are connected to a
network analyzer [13]. The unknown material is placed in the line-of-sight connecting the two
antennas. In order to extract the dielectric properties of the material, the same structure is
simulated using commercial full wave electromagnetic simulator software. In the full wave
simulation, the dielectric properties of the material are swept in a wide range to match the
simulated and measured results. As a result, a set of εr and tanδ values will be found in
simulation which give the same transmission coefficient as the measured one. This will be
reported as the dielectric properties of the material at that particular measurement temperature. In
order to achieve the material’s properties at other temperatures, the sample needs to be heated
and the same process should be repeated. The main advantage associated with this method is the
physical separation of the measurement ports and antennas with the sample under test, which
makes this method attractive for high temperature measurements where proper protection of the
measurement devices is a challenge. The drawback associated with the free space method is the
requirement of large sample sizes in order to achieve reasonably high measurement accuracy.
This makes characterization of small samples almost impossible using this method. This problem
becomes more critical for applications in need of tiny material sizes for miniaturized sensor
development.
The other transmission based method is parallel-plate or capacitive characterization scheme [14].
As shown in Fig. 1.2(b), the capacitance measured across two parallel plates changes as a
dielectric material is inserted between the plates. In this approach, the unknown material will be
7
placed between two metallic and parallel plates and the capacitance will be measured at each
temperature. Using basic circuit theory, the corresponding dielectric constant of the material can
be calculated at each temperature. This method is very well suited for thin and flat materials and
the involved calculations and data processing are quite simple. On the other hand, the maximum
characterization frequency and its accuracy are limited. In addition, the parallel plate method
lacks the necessary accuracy in the loss measurement.
Figure 1.3: Different reflection based material characterization methods which can be used for
high temperature material characterization; (a) rectangular/circular waveguide, and (b) coaxial
probe.
The last class of transmission based material characterization methods consists of a waveguide
which is partially loaded with the unknown material [15]. The two common configurations used
in this method are the rectangular/circular waveguides and coaxial lines as illustrated in Fig.
Hot
zone
Sample
Under Test
Ne
two
rk A
na
lyze
r
Rectangular
or Circular
Waveguide
Short
Circuit
Hot
zone
Sample
Under Test
Ne
two
rk A
na
lyze
r
Coaxial
Waveguide
(a) (b)
8
1.2(c) and (d). As for the previous two methods, the transmission coefficient measured at the two
ports will be affected by the presence of the material under study. Again, full wave simulations
will be used to extract the material properties. This characterization method is simple to
implement and analyze. However, it requires precise machining of the material to obtain the
exact cross-sectional dimension of the waveguides. In addition, the small air gaps occurring
between the material and the waveguide cause measurement uncertainties.
Beyond the transmission based characterizations, the next category of the material
characterization methods is based on measuring the reflection coefficient of a waveguide loaded
with the unknown material. There are two methods to be discussed in this category. In the first,
the waveguide will be closed at one end with an electrically conducting plate and the unknown
sample will be placed right before the short circuit as shown in Fig. 1.3(a) [16]. This method is
similar in principle to the waveguide transmission method illustrated in Fig. 1.2(c) and has the
same benefits and limitations, except for the fact that there will be only one measurement port.
However existence of only one measurement port makes this configuration attractive. The other
reflection based material characterization method is shown in Fig. 1.3(b) and is referred to as the
coaxial probe method [17]. In this method, the sample is placed at one end of an open-ended
coaxial line. However, there is no short circuit as previously described for the
rectangular/circular waveguide case and the sample is suspended in free space. There are no
strict requirements for the sample’s dimensions when using this method and it provides more
flexibility in characterizing samples with various shapes. However, the air gap between the
sample and the probe can be the source for significant measurement errors.
9
Figure 1.4: Different resonator based material characterization methods which can be used for
high temperature material characterization; (a) rectangular sample with the same cross-section
dimension, and (b) cylindrical sample with arbitrary dimensions.
The last category of material characterization methods is illustrated in Fig. 1.4 and is based on a
cavity resonator [18][19]. This is the most accurate characterization method which will be
described in Chapter 3. This technique is based on a metallic cavity resonator which is coupled
to the access ports enabling measurement of its resonant frequency and quality factor. The
unknown material is then placed inside the cavity which modifies the aforementioned parameters.
Full wave simulations are again used for the extraction of material properties. Being accurate,
this method is very well suited for applications in need of precise material properties such as
wireless sensors. However, realization of this technique at high temperatures is challenging. In
addition, the other drawback associated with this method is its single frequency performance. In
Hot
zone
Sample
Under Test
Network Analyzer
Cavity Hot
zone
Sample
Under Test
Network Analyzer
Cavity
(a) (b)
10
other words, the measurement frequency is predefined by the cavity dimensions and cannot be
easily tailored.
1.3 Overview of Dissertation
This dissertation presents the high temperature material characterization method and the design
and test of three types of wireless high temperature sensors. Chapter 2 presents the fabrication
process of the SiCN materials and the dielectric property at low frequencies. In Chapter 3, a
waveguide cavity method is used to characterize the material properties at the room temperature
with very high accuracy. The measured dielectric constant of SiCN material using this
waveguide cavity method is used to estimate the resonant frequency of the resonator for high-
temperature measurement and therefore is used to determine the dimensions of the TRL
calibration kit. Chapter 4 presents the design and fabrication of high-temperature characterization
setup as well as the TRL calibration kit. The measurement results of the dielectric properties of
SiCN ceramics are also shown in the same chapter. In Chapter 5, three different types of sensors,
corresponding to different antenna configurations, are designed and the prototypes are fabricated
and tested. Finally, the conclusions and future works are described in Chapter 6.
11
CHAPTER 2 HIGH TEMPERATURE MATERIALS
2.1 SiCN Ceramic Material
SiCN ceramic materials are polymer-derived ceramics (PDCs), which are a new class of
materials synthesized by thermal decomposition of polymeric precursors. PDCs possess
excellent thermo-mechanical properties up to very high temperatures (> 1500oC), such as
excellent thermal stability, high oxidation/corrosion resistance, and creep resistance, thus making
them suitable for high-temperature applications [20][21][22][23][24][25][26][27][28][29][30].
They are potential candidates to develop high-temperature MEMS sensors for turbine engines.
2.2 Fabrication Procedure
The amorphous SiCN ceramics can be fabricated through the thermal decomposition of a
commercial liquid-phase precursor using the technique reported previously in [31], as shown in
Figure 2.1 and Figure 2.2. Precursor is a compound that participates in a chemical reaction that
produces other compounds. The one used in this paper is “VL20” which was purchased from
KiOH Corporation. First, the precursor is solidified by heat treatment at 250oC for 1 hour, and
then being heated at 350oC for 4 hours in a flow of ultra-high-purity nitrogen. The solids
obtained from the previous step are then crushed and ball-milled into fine powders using a high-
energy ball milling machine. Finally, the powders are compressed in the shape of disc and then
pyrolyzed at 1400oC or higher for 4 hours in a tube furnace with the ultra-high-purity nitrogen
flowing. The scanning electron microscope (SEM) image shown in Fig. 1.3 reveals the typical
microstructure of the pyrolyzed SiCN samples. In order to achieve better dielectric property
12
(high dielectric constant, low loss tangent) of SiCN materials, boron(B)-doped SiCN ceramics
with the different ratios of Si to B are fabricated and characterized in this dissertation.
Figure 2.1:Process flow of SiCN ceramics in schematics
Figure 2.2: Process flow of SiCN ceramics in photos.
1hr@250oC
4hr@350oC
Ball milling
4hr@1400oC
In N2
Co
mp
ress
H
D
13
Figure 2.3:Scanning electron microscope (SEM) images of the SiCN material at 500× zoom level.
2.3 Dielectric Property of SiCN at DC and Low Frequencies
Figure 2.4 shows the temperature dependence of dielectric constant and tangent loss of sample
SiCN (sintered at 1200oC) measured at frequency of 1, 10, and 100 KHz over temperature range
of 20 ~ 675ºC, respectively [32]. In general, dielectric constant increases with temperature.
14
Figure 2.4: Temperature dependence of the dielectric constant of sample SiCN (sintered at
1200oC) measured at frequency of 1, 10, and 100 KHz over temperature range of 20 ~ 675 ºC
[32].
2.4 Conclusion
SiCN ceramic materials have been demonstrated to be very stable and corrosion resistant for
high temperature, which make them good candidates for high-temperature sensing applications.
It was also discovered that their dielectric constant increases with temperature below 1 MHz.
However, the proposed wireless passive sensing mechanism requires the use of microwave
frequencies for reliable temperature sensing without the need of wire connections and packaging.
Therefore, it is highly desired to characterize the dielectric properties of SiCN materials at
microwave frequencies which will be used as the carrier frequencies in wireless sensing.
15
CHAPTER 3 ROOM TEMPERATURE CHARACTERIZATION
3.1 Introduction
As one of the important steps in high-temperature characterization, material properties of SiCN
ceramics need to be characterized at room temperature first with high accuracy. There are two
main reasons for this study. First, the measurement results at room temperature can provide a
reference for the high-temperature characterization. It will be shown later that the measured
dielectric properties agree very well for both measurement methods at room temperature. This
ensures the validity of the results at other temperatures as well. Second, the room-temperature
dielectric properties of SiCN materials will be required to design a custom-made TRL calibration
kit to improve the accuracy for the high-temperature characterization.
There are several well-known techniques for characterizing the dielectric constant (εr) and loss
tangent (tanδ) including the transmission line methods, free-space methods, and cavity methods
[19][33][34][35]. Transmission line methods are able to characterize the material properties over
a very wide frequency range. However, the loss tangent measurement is usually not accurate
enough. Free-space methods require a large sample to avoid edge diffraction and are usually
used at millimeter-wave frequencies and above. Perturbation methods using waveguide cavities
are easy to derive the dielectric constant and loss tangent. Nevertheless, the limited amount of
energy storage within the test samples could cause a poor estimate of the loss tangent,
particularly for low-loss-tangent materials. A fully-filled cavity is ideal to characterize the
material properties with high accuracy, but it is very difficult to prepare the high-temperature
ceramic material samples to fit in the cavity exactly. The resulting air gap between the sample
16
and cavity can greatly compromise the measurement accuracy. To address the need of
characterizing low-loss-tangent materials with high accuracy, a cavity method with the assist of
full-wave simulations was presented in [19] and was able to achieve very high accuracy with
excellent repeatability. Loss tangent as low as 810-5
has been measured and independently
verified by National Institute of Standards and Technology (NIST). In this work, these high-
temperature ceramic materials exhibit both low dielectric constant and low loss tangent in
contrast to the high dielectric constant and low loss tangent of the titania ceramics studied in [19].
As a result, dielectric resonator modes used in [19] are not visible for the materials studied herein.
An alternative resonant mode is used to characterize the dielectric constant and loss tangent in
this chapter.
3.2 X-band Adapter Cavity Characterization
3.2.1 Characterization Setup
The measurement setup is illustrated in Figure 3.1. Ceramic samples are placed inside a metallic
waveguide cavity. The dimensions of the waveguide cavity are 53.71 mm by 22.78 mm by 10.08
mm. A Rogers Duroid 5880 spacer (11.20 mm in diameter and 0.8 mm in height, εr = 2.2, tanδ =
0.0009) is used to elevate the ceramic samples. It was found in [19] that the sensitivity of the
resonant frequency and unloaded Q factor (QU) caused by the air gap between the ceramic
samples and the bottom of the cavity is very high if no spacer is used. This cavity is weakly
coupled by two open-ended coax connectors [36].
17
Ceramic
Coax Connecters
Duroid Spacer
Figure 3.1: Measurement setup for SiCN sample.
First, the resonant frequency and QU of the lowest mode of the empty waveguide cavity are
measured to be 7.127 GHz and 3,216, respectively. Therefore, the effective conductivity of the
waveguide cavity wall is found to be 1.178107 S/m by performing full-wave Ansoft High
Frequency Structure Simulator (HFSS) simulations. This effective conductivity will be used in
the simulations of the loaded cavity. In [19], dielectric resonator (DR) modes were used since
these modes maximize the amount of energy storage inside the test samples, consequently
maximizing the measurement accuracy. The resonant peaks for a SiCN sample are measured
using an Agilent performance network analyzer (N5230A) and shown in Figure 3.1. To match
these measured resonant peaks, a parametric sweep of dielectric constants is performed in HFSS.
A good agreement is found between simulation and measurement results. In order to identify the
resonant modes of these peaks, electric field distributions of different modes are simulated in
HFSS and presented in Figure 3.2 using the dielectric constant found in the parametric sweep.
Both driven mode and eigen mode simulations are performed. It is found that the first DR mode
18
occurs at 11.684 GHz in the eigen mode simulation. However, this mode is not visible in both
driven mode simulations and measured results. This is due to the fact that DR modes can well
confine the energy within the sample, causing a very low energy transfer at these modes. In
addition, as observed in Figure 3.1, there are two strong transmission peaks at 10.050 GHz and
11.938 GHz, respectively. The transmission peak of the DR mode at 11.684 GHz could possibly
be inundated by the two adjacent strong peaks.
Figure 3.2: Measurement result for SiCN sample.
In order to minimize the uncertainty of characterizing the permittivity of the ceramic materials,
maximal energy storage inside ceramic samples is desired. Therefore, the volume fraction (Vf) of
each resonant mode needs to be identified. Vf is the percentage of energy stored inside the
samples [37]. The waveguide cavity mode with the highest volume fraction (Vf) of the electric
6 7 8 9 10 11 12 13 14 15 16
-70
-60
-50
-40
-30
-20
-10
0
S21
(dB
)
Frequency (GHz)
Measured
Simulated
19
field inside the sample was chosen to achieve the best measurement accuracy. The following
relation exists between Vf and different quality factors:
(3.1)
where QU is the unloaded Q factor of the cavity loaded with the sample, Qd is the inverse of loss
tangent of the material to be characterized, Qs is the inverse of the loss tangent of the materials
for the spacer, Qm corresponds to the metal loss of the cavity, Vfd is the volume fraction inside the
measured sample, and Vfs is the volume fraction inside the spacer. An eigen mode simulation
with just the loss tangent of the ceramic sample specified is performed to give a QU. The inverse
of the product of this QU and the specified loss tangent is equal to Vf.
Driven Mode Simulation Eigen Mode Simulation
6.198 GHz
6.212 GHz
8.284 GHz 8.298 GHz
20
Figure 3.3: E field of the first five resonant modes for the SiCN sample
The geometrical dimensions of the SiCN sample are 4.64 mm in diameter and 5.97 mm in height.
The Vf of the first four resonant modes for the SiCN sample are listed in Table 3.1. It is apparent
that the first mode of both samples has the largest Vf. As a result, the first mode is used for the
analysis in the next section.
10.050 GHz 10.070 GHz
NO Visible Coupling
11.684 GHz
11.938 GHz
11.958 GHz
21
Table 3.1: Vf of different modes for SiCN sample
3.2.2 Measurement Results
The measured resonant frequencies and QU of the SiCN sample are listed in Table 3.2. Three
independent measurements are performed to verify the measurement repeatability. The εr and
tanδ of the SiCN sample are obtained by parametric sweeps in HFSS to match the measurement
results with the material properties of the waveguide wall and spacer specified.
The average and standard deviation for both εr and tanδ are shown in Table 3.3. The calculated εr
of the SiCN sample is 4.358 with a 0.58% standard deviation, demonstrating a very consistent
measurement. The tanδ is found to be 5.26 10-3
with a standard deviation of 5.72%, which is
very good since a larger variation in the Q measurement is normally expected.. A summary of
the measured material properties of the SiCN ceramic material is listed in Table 3.
The accurate estimation of the dielectric constants of these high-temperature ceramic materials is
critical for the design of temperature sensors for turbine engines. The low loss tangents of both
Mode Freq. (GHz) Vf
1 6.212 22.4%
2 8.298 5.1%
3 10.070 11.9%
4 11.958 19.2%
5
22
materials are important to achieve high-Q resonances which ultimately determine the sensitivity
and sensing range of the sensor systems.
Table 3.2: Summary of results for SiCN
Table 3.3: Average and standard deviation (ζ) of εr and tanδ
3.3 Ku-band Adapter Cavity Characterization
This section improves the waveguide cavity characterization method proposed in previous
section by further increasing Vfd and optimizing the spacer design to achieve better measurement
accuracy. In [19], although two H-band coax-to-waveguide adaptors were used to form the
cavity for characterization, Vfd was still large because of the high dielectric constant of test
samples. In the previous section, two X-band coax-to-waveguide adaptors were used. In this
section, two Ku-band adaptors are employed to increase Vfd from 22.4% to 33.4% for the same
No. Resonant Freq.(GHz) QU εr tanδ (10-3
)
SiCN
1 6.218 691.9 4.387 4.91
2 6.225 668.2 4.340 5.41
3 6.224 664.9 4.347 5.45
Material Avg. εr ζ (εr) Avg. tanδ (10-3
) ζ (tanδ)
SiCN 4.358 0.58% 5.26 5.72%
23
sample material and size. There are different electric field distributions for different resonance
modes. This results in Vf different for each resonant mode. The dielectric constants of the test
sample also affect Vf. For the same size of the samples, a higher dielectric contrast results in a
higher Vf since the field is more tightly constrained within the test dielectric sample. The Ku band
adaptor cavity is smaller than X-band adaptor cavity, so for the same resonant mode there is
more electric field energy in the test sample. In addition, the resonant frequency corresponding to
the waveguide cavity mode of interest, increases from 6.22 to 7.899 GHz, which is much closer
to the resonant frequency of the dielectrically-loaded cavity resonator used for the high-
temperature characterization. To reduce the measurement uncertainty caused by the spacer, a
hollow ring shape is adopted to minimize Vfs while maintaining the mechanical support of the
ceramic sample. Since SiCN materials exhibit dielectric constants around 3-5, the electric field
confinement inside the SiCN sample is not as concentrated as in the Titania cases. The dielectric
constant of porous Titania varies from 12 to 90 [19]. Because of the ring spacer’s hollow
structure, less electric field energy is concentrated in it, compared with the solid spacer, and the
Vf of the ring spacer is smaller. Therefore, the Vf of the test sample is larger for the ring spacer
case and the measurement uncertainty decreases. In addition, according to the datasheet from the
vendor, the dielectric constant of materials used for spacer is in the range of from 2.18 to 2.22,
however, in the simulation, the mean value of the dielectric constant, 2.2, was used. Therefore, it
involves the error for the characterization. The spacer in the shape of ring reduces the size of the
spacer compared with a solid spacer, so the measurement uncertainty is reduced, too. Therefore,
the adverse effects from a solid spacer are more pronounced compared with the Titania cases.
This ring-shaped spacer design can provide near-air supporting structure underneath the material
24
sample, and therefore represents an optimized design. It should be noted that RT/Duroid 5880 is
chosen here again for its low dielectric constant of 2.2 and low loss tangent of 0.0009.
3.3.1 Characterization Setup
Figure 3.4: Illustration of the waveguide cavity method. For 5880 measurement, Ds=14; Hs= 3.19;
Di_r=8; Do_r=9; Hr=3.19. For SiCN measurement, Ds=9.29 Hs= 6.26; Di_r=6; Do_r=7.08; Hr=0.08
(unit: mm).
Figure 3.4 illustrates the characterization setup, which is consisted of a metallic waveguide
cavity loaded with the to-be-measured sample. The corresponding measurement setup is shown
in Fig. 3.5. As shown in Figure. 3.5, two Ku-band coax-to-waveguide adaptors are assembled
together to form a waveguide cavity with dimensions of 52.46×15.74×7.86 mm3. The
dimensions of the samples and spacers are shown in Figure 3.4. The cavity is weakly excited by
two coaxial probes. The cylindrical sample is elevated using a ring-shaped spacer. The dielectric
constant and loss tangent of the material under study can be extracted from the measured
Spacer Sample
Connector
Di_r
Do_r
Hr
DsHs
Y
Z
X
25
resonant frequency fr_w and unloaded Q factor QU_w of the cavity containing the sample, with the
assist of Ansoft High Frequency Structure Simulator (HFSS) full-wave simulation software
package. The eigen-mode simulation is used for the dielectric property extraction. The number of
tetrahedra for meshing the simulated structure is more than 40,000 and the ΔS for the
convergence criteria is less than 0.0005.
Figure 3.5: Waveguide cavity method setup. View of (a) the half cavity at the assembly plane
and (b) the whole cavity after assembly.
The measurements of two materials, i.e. RT/Duroid 5880 and SiCN, are performed using this
improved waveguide cavity method at room temperature. The size of SiCN samples is restricted
by the molds used in fabrication and is also affected by the different fabrication process, such as
sintering temperature, time, and other experimental parameters. The dimensions of the SiCN
samples for characterization are about 10 mm in diameter and 5 mm in height. The ring spacer
Spacer
Sample
(a) (b)
26
should be large enough to elevate the test sample in the middle of the cavity in the z direction
and the sidewall should be as thin as possible to reduce the electric field concentration inside the
spacer. The size of RT/Duroid 5880 samples as shown in Fig. 3 is in the range of the size of
SiCN sample. Its Vf is smaller than SiCN case. Therefore, the accuracy for the SiCN
characterization is not less than that for RT/Duroid5880 characterization.
3.3.2 Measurement Results
To verify the measurement accuracy of the improved waveguide cavity method, RT/Duroid 5880
material is characterized, which represents a challenging case for its low dielectric constant of
2.2±0.02 and low loss tangent of 0.0009. Ten independent measurements are performed to
identify the measurement consistency. The average extracted εr and tanδ are 2.18 and 0.0008,
which match the nominal values very well. The standard deviations for εr and tanδ are 0.10% and
3.61%, respectively, which are much better than 0.58% and 5.72% presented in the previous
section. The Type A experimental uncertainties for for εr and tanδ are 0.001 and 0.00001,
respectively, based on the statistical analysis in [38]. This is primarily attributed to the increased
Vfd and improved spacer design which corresponds to reduced Vfs.
27
Figure 3.6: S21 results of 10 measurements for SiCN sample characterization.
Table 3.4: Measured fr_w, QU_W and extracted εr, tanδ with standard deviations (ζ) and
uncertainty (su) using waveguide cavity method at room temperature
The same measurement and simulation process is performed for the SiCN material. The
dimensions of the SiCN sample and spacer are shown in Fig.3.4. Again, 10 independent
measurements are performed. Fig. 3.6 shows S21 results for these measurements. The average
extracted εr and tanδ are 3.655 and 0.00408, respectively. The standard deviations for εr and tanδ
Measured
Sample fr_w (GHz) (ζ) (su) εr (ζ) (su) QU_w (ζ) (su) tanδ (ζ) (su)
5880 9.1126 (0.01%)
(0.0004)
2.182 (0.13%)
(0.001)
2292
(1.20%) (11)
0.00082 (3.61%)
(0.00001)
SiCN 7.8989 (0.03%)
(0.0006)
3.655 (0.13%)
(0.001)
540 (2.15%)
(3)
0.00408(2.16%)
(0.00002)
7.80 7.85 7.90 7.95 8.00-70
-65
-60
-55
-50
-45
-40 Meas.1
Meas.2
Meas.3
Meas.4
Meas.5
Meas.6
Meas.7
Meas.8
Meas.9
Meas.10
S2
1 (
dB
)
Frequency (GHz)
28
are 0.13% and 2.16%, respectively, which are again much better than the results in X-band
cavity characterization. The Type A experimental uncertainties for deviations for εr and tanδ are
0.001 and 0.00002, respectively,The measurement results as well as extracted complex
permittivity for both materials are summarized in Table 3.4. It is noted that this waveguide cavity
approach can provide very high measurement accuracy; however, it cannot be easily extended
for high-temperature measurements. A new characterization method which is specifically
designed for high-temperature characterization of dielectric materials at microwave frequencies
will be presented in the next section. This method requires relatively small sample sizes and can
be tailored to different microwave frequencies with high measurement accuracy.
3.4 Conclusion
The dielectric properties of SiCN ceramic materials have been characterized at microwave
frequencies using an accurate measurement technique based on cavity resonator and full-wave
simulations. This material property information serves as a basis to develop wireless passive
ceramic MEMS high-temperature sensors for turbine engines. It is noted that this waveguide
cavity approach can provide very high measurement accuracy; however, it cannot be easily
extended for high-temperature measurements. A new characterization method which is
specifically designed for high-temperature characterization of dielectric materials at microwave
frequencies will be presented in the next chapter. This method requires relatively small sample
sizes and can be tailored to different microwave frequencies with high measurement accuracy.
29
CHAPTER 4 HIGH TEMPERTURE CHARACTERIZATION
4.1 Introduction
SiCN is stable and corrosion resistant at high temperatures with temperature-dependent
permittivity at low frequecies. These advantages make SiCN a suitable candidate for high
temperature wireless sensor development. However, the wireless passive sensing mechanisms
require the use of microwave frequencies for reliable temperature sensing without the need of
wire connections and packaging. Therefore, it is necessary to characterize the dielectric
properties of SiCN materials at microwave frequencies and high temperatures, which will be
used as the carrier frequency in wireless sensing. In addition, this measurement setup can be
easily adapted for wireless passive sensing.
Figure 4.1: Flowchart of the high temperature characterization mechanism.
Temperature
variation
εr and tanδ of SiCN
material change
Resonant frequency and
Qu of the cavity resonator
change
Scattering parameters
change
Full wave simulations
extract corresponding
εr and tanδ
Rogers RT/duriod 5880 SiCN
r
rf
tanδ
S
Qu
30
4.2 Proposed High-Temperature Characterization Method
Fig. 4.1 illustrates the mechanism behind the high-temperature characterization setup which is
shown in Fig. 4.2. The dielectric material under study is first metalized to form a dielectrically-
loaded cavity resonator. The material could take any shape. In this paper, a cylindrical form is
used. Two opening slots are required to allow the transmission of microwave signal through the
resonator. The resonator is then placed over short-ended CPW lines. The CPW line and resonator
are weakly coupled through the aforementioned slots. The resonator and CPW line are simply in
contact without glue. Because the contacting surfaces of the resonator and CPW line are very flat,
this connection can still achieve the desired coupling level. Moreover, only fr and Qu which are
not very sensitive to the level of S21 are used for the extraction of the dielectric properties of the
test materials. SMA connectors are connected to the CPW lines. The substrate material used for
the CPW line is alumina. Extracting dielectric properties is the inverse process. First, The
scattering parameters of the described setup are measured at various temperatures. Then, the
resonant frequencies and unloaded Q factors of the resonator with the test sample can be
extracted from the measured S results. The next step is to use the full simulation tool HFSS to
extract the dielectric constant and loss tangent of the test sample. The parametric sweeps of
dielectric constants and loss tangent of the test sample are performed using HFSS to match the
measured resonant frequency and unloaded Q. Finally, when the simulated resonant frequency
and unloaded Q are the same to the measurement results, the dielectric constant and loss tangent
of the material under study are extracted.
31
Figure 4.2: Illustration of the high temperature characterization method
The resonant frequency (fr) and Q of the TM010 mode inside the dielectrically-loaded cavity
resonator can be calculated using the following equations [39]:
(4.1)
(4.2)
(4.3)
(4.4)
End Launch
Connector
CPW line
Resonator
Coupling Slot
High Temperature
Zone
32
where ε0 and μ0 are the permittivity and permeability of free space. χ01 is the first root of Bessel
function of the first kind J0(z), i.e. 2.2048, and ζ is the bulk conductivity of the metal. In this
study, gold is used for metallization purpose with ζ = 4.1×107 S/m. D and H represent the
diameter and height of the cavity. εr and tanδ are the dielectric constant and loss tangent of the
dielectric material, respectively.
Figure 4.3: The proposed high temperature material characterization setup.
Network Analyzer
Resonator
Furnace
Temperature
controller
DC power
supply
Power
cable
Thermocouple
115 Volt
60 Hz Line
Source
CPW line SMA ConnectorCoaxial line
33
In Chapter 1, different material characterization techniques have been described. It was shown
that some of the methods benefit from being simple and easy to implement, while others provide
higher accuracy. In addition, in order to use any of those methods to characterize materials at
extremely high temperatures for turbine applications, careful considerations need to be made.
The “hot zone” illustrated in Figs. 1.2 to 1.4 is an abstract concept which needs to be properly
realized. The physical realization of this hot zone should not overheat the measurement facilities
including the connectors, cables and the network analyzer. In addition, stable temperature at the
vicinity of the sample is required. Therefore, the heating element in the measurement apparatus
should occupy a substantial portion of the furnace to mitigate fast temperature fluctuations. The
other issue to be considered is the temperature control loop. There should be direct access to the
sample in order to monitor its real temperature during the measurement. This requires the
measurement setup to provide access to the sample under test while maintaining a stable local
temperature, which is a challenging task, as such access generally exposes the measurement
environment to the external environment.
To achieve the aforementioned requirements at the same time, the characterization setup is
presented as illustrated in Fig. 4.3. A CPW transmission line is loaded with a cavity resonator in
the middle and passes through a cylindrical ceramic furnace. There are a number of advantages
associated with this setup. Using a transmission line alleviates any reliance on fully-confined
measurement chambers, allowing for realization of a relatively large hot zone compared with the
sample size. In addition, the open ends of the furnace provide proper access to the sample in
order to monitor the local temperature at the resonator using a thermocouple. Use of CPW
configuration as the transmission line enables real time observation of the sample which is not
34
possible in the closed environment of a rectangular/circular waveguide or coaxial line. In
addition, the proposed configuration provides reasonable high temperature gradient outside the
hot zone to protect the measurement facilities.
Figure 4.4: Details of the proposed material characterization mechanism including a CPW line
loaded with a cavity resonator.
Fig. 4.4 shows more details about the presented characterization method. The CPW line
consisting of the dielectric layer, signal and ground lines is illustrated. The cavity resonator is
formed by metalizing the material under test. Two small slots are opened at the sidewalls of the
cavity to enable coupling of the signal into and out of the resonator. The resonant frequency and
quality factor of the cavity can be consequently measured at the two ports of the CPW line.
These parameters are functions of the dielectric properties of the unknown material inside the
cavity at the operating temperature. After measuring the resonant frequency and quality factor of
the resonator, full wave simulations are employed to extract the corresponding dielectric
CPW
LineResonator
Coupling Slots
Top View
Side View
Ground
GroundSignal
35
properties of the material under study. The materials to be characterized here is silicon carbon
nitride (SiCN) without and without boron doping (SiBCN) developed at UCF. SiCN is stable and
corrosion resistant at high temperatures with temperature-dependent permittivity. These
advantages make SiCN a suitable candidate for high temperature wireless sensor development.
However, before being able to utilize SiCN in the sensor development, the material needs to be
characterized at different temperatures. To do so, two different metallization processes are
implemented for the measurements with the highest temperatures of 500oC and 1300
oC,
respectively.
4.3 Characterization Up to 500oC
4.3.1 Measurement Setup Design
It is noted that to design a dielectrically-loaded cavity resonator, dielectric properties of the
material need to be identified first. The measured SiCN material properties using the cavity
method presented in the previous section, i.e. εr = 3.65 and tanδ = 0.005, are employed here to
achieve this goal. fr and QU for the enclosed cavity are calculated to be 12.77 GHz and 189.2,
respectively. However, as the cavity is coupled to the CPW lines through the slots, its fr and QU
will change. In addition, the temperature not only modifies the fr and QU but also affects the
transmission level. Therefore, a parametric study to show the effect of coupling slot dimensions
on fr, QU, and transmission level is necessary. The full-wave simulations are performed with
reference planes at the coaxial connectors.
36
Figure 4.5:: Dimensions of (a) CPW line. (b) fabricated SiCN cavity resonator and (c) connector.
S = 1.5; G = 0.4; W = 10; T = 0.635; d = 70; D = 9.41; H = 5.01; H1 = 1; W1 = 3; L1 = 9.5; L2 =
7.92; D1 = 3.9; D2 = 0.76; S1 = 0.81. (unit: mm).
Fig. 4.6 (a) shows the fr corresponding to the peak in the simulated S21 for different slot
dimensions. The fr changes from 12.23 to 12.87 GHz for different combinations of slot widths
(2-4 mm) and heights (0.2-2 mm). It is observed that as the slot width or height increases, the fr
decreases. Additionally, higher sensitivity to the slot size is observed at larger dimensions.
Qu of the cavity can be extracted from the measured loaded Q (QL) using [40]:
(4.5)
where is the value of S21 in linear at the resonant frequency.
As seen in Fig. 4.6(b), smaller Qu is achieved for larger slot sizes. Again, higher sensitivity to the
slot size is observed at larger dimensions.
WG
S
d D d
D
H
TW1
H1
(a)
(b)
L1
L2
D1D2
S1
(c)
37
Figure 4.6: Effect of the coupling slot dimensions on (a) resonant frequency, (b) unloaded Q, and
(c) transmission level.
Fig. 4.6(c) shows the peak level of the simulated S21 for different slot dimensions. In general,
larger slot dimensions produce higher transmission levels. However, this trend does not hold
when W1 is above 3.5 mm.
Since the transmission level decrease with rising temperatures which will be shown later,
maximum transmission level is desired at room temperature. From Fig. 6(c), it is observed that
the slot width should be more than 3 mm to have a transmission level around -20 dB. Fig. 8(b)
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0-36
-34
-32
-30
-28
-26
-24
-22
-20
-18
S21 (
dB
)
H1 (mm)
W1=2mm
W1=2.5mm
W1=3mm
W1=3.5mm
W1=4mm
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.012.2
12.3
12.4
12.5
12.6
12.7
12.8
12.9
f r(GH
z)
H1(mm)
W1=2mm
W1=2.5mm
W1=3mm
W1=3.5mm
W1=4mm
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.040
60
80
100
120
140
160
180
200
QU
H1(mm)
W1=2mm
W1=2.5mm
W1=3mm
W1=3.5mm
W1=4mm
(a) (b)
(c)
38
suggests smaller width for higher Qu. Therefore, W1 and H1 are chosen to be 3 and 1 mm,
respectively, by trading off between performance and sensitivity. The simulated fr and Qu of the
resonator as shown in Fig. 4.5 with the aforementioned slot dimensions are 12.70 GHz and 151.8,
respectively.
4.3.2 TRL Calibration Kit
In order to keep the connectors at low temperatures without damaging the cables connected to
the network analyzer when the sample is heated to high temperatures, the length of the alumina
board should be above 14 cm, which is based on the dimensions of the furnace used for
characterization. Short-Open-Load-Through (SOLT) calibration was carried out at the reference
planes of the SMA connectors. The measurement results revealed undesirable ripples and
excessive losses which came from the long transmission lines and the transitions between the
connectors and alumina board. In order to remove the undesirable effects from the long CPW
line and achieve better accuracy, custom-made TRL calibration standards are designed and
fabricated, similar to the approach used by NASA for high-temperature characterization on
Sapphire and Alumina substrates [41].
39
Figure 4.7:Designed TRL calibration kit at 12.4 GHz. (a) Through, (b) Reflect, and (c) Line. d =
70, L = 2.57. (Dimensions in mm)
Figure 4.8: Fabricated TRL calibration standards.
(a) (b)
2*d+L
2*d d
(c)
40
As shown in Fig. 4.7, a TRL calibration kit is designed at 12.4 GHz, which corresponds to the fr
of the cavity resonator at around 500oC as shown in the next section. The substrate is ADS-96R
alumina with a thickness of 0.635 mm. The widths of the center conductor and slot are 1.5 and
0.4 mm, respectively. The electrical length of L is typically designed at 90o but can be between
20o and 160
o [42]. Due to the restrictions from the fabrication equipment for patterning and
cutting, L is chosen to be 2.57 mm, corresponding to an electrical length of 62o which is around
1/6 λg. λg is the guide wavelength of the CPW line at 12.4 GHz. The fabricated TRL calibration
kit is shown in Fig. 4.8.
4.3.3 Metallization
The metallization of the SiCN sample is performed in two steps. First, a 0.3-μm-thick Cr layer is
evaporated on the sample. It is noted that due to the directive nature of evaporation, five
evaporations at different angles are needed to cover the entire surface of the sample. Then, a 7-
μm-thick Au layer is electroplated to minimize the conductor loss due to the skin depth effect.
The temperature of the gold plating solution and the voltage cross the anode and cathode has
been optimized through many experiments to realize uniform coverage of Au. The solution is
stirred and heated in the range of 65±2oC using a stirring hot plate. To ensure high plating quality,
the voltage should be less than 1.5 volts and the current density passing through the surface of
the sample should be kept less than 4 mA/in2. Fig. 4.9 shows the variation of voltage and current
density during the gold electroplating process. In Stage 1, since the resistance of the 0.3-μm-
thick Cr seed layer is relatively high, the voltage is slowly ramped up from 0.5 to 1.5 V when the
current density increases from 0 to 4 mA/in2 in a nonlinear fashion. t0 depends on the sample
size and is approximately 10 min in this study. To keep the current density at 4 mA/in2 in Stage 2,
41
the voltage needs to be decreased as shown in Fig. 4.9 since the resistance of the metal layer
becomes smaller as the gold layer gets thicker. The coupling slots are formed by taping during
the evaporation. The final dimensions of the two coupling slots are measured to be 3.1 × 0.6 mm2,
and 2.7 × 0.8 mm2, respectively, due to fabrication tolerances.
Figure 4.9: Voltage and current density variations during gold electroplating.
The CPW lines on the alumina board are photolithographically defined and metalized using the
same procedure as for the sample.
4.3.4 High Temperature Measurement Setup
The measurement is performed from 25 to 500oC, which is limited by the choice of metals in this
study. SiCN ceramic materials have been demonstrated to be thermally and mechanically stable
in corrosive gases and at temperatures up to 1500oC.
00.0
0.5
1.0
1.5
2.0
Stage 2
Voltage
Current Density
Time (min)
Vo
lta
ge
(V
)
0
2
4
Stage 1
Cu
rren
t De
ns
ity (m
A/in
2)
t0
42
Figure 4.10: (a) The connector before soldering. (b) The furnace containing high-temperature
measurement setup connected to the RF cables. (c) Sample over CPW line outside the furnace. (d)
Sample and CPW line inside the furnace. (e) Sample inside the furnace with thermal isolation at
the open ends of the furnace.
The high-temperature measurement setup is shown in Fig. 4.10. To create a high-temperature
zone with precise temperature control, a cylindrical furnace (Robust Radiator from Micropyretics
Heaters International Inc., Model number RHUL-MP1125-4) is used. This furnace has a
(a)
(d)
K-type
thermocouple Resonator
CPW line
(b)
(e)
(c) (a)
43
cylindrical hot zone of 25.4 mm in diameter and 50.8 mm in length, which can be heated up to
1500oC with ±2
oC accuracy using a closed-loop feedback temperature controller. In the
measurement, a programmable DC power supply (GenesysTM
1500W from TDK-Lambda
Americas Inc., Model number GEN 12.5-120) and a temperature controller (Omega CN96211TR)
are used to realize precisely temperature control inside the furnace. The alumina board is placed
inside the furnace with both ports extending outside the furnace. The metalized SiCN resonator
is located in the middle of the hot zone.
To avoid ventilation and temperature instability, the two open ends of the furnace are sealed
using Alumina Silica Ceramic Fiber extreme-temperature sheeting (Al2O3 Wool) which can
withstand up to 1260oC. With the sealing, the temperature at the open ends of the CPW lines
outside the furnace is measured to be around 120oC when the hot zone is heated to 500
oC.
Therefore, SMA connectors remain intact during the entire measurement procedure.
4.3.5 Measurements Results
TRL calibration is performed using an Agilent Performance Network Analyzer (PNA) N5230A
at every temperature point where the sample is measured. IF-bandwidth is 1.5 KHz, spanned
frequency range from 12 GHz to 12.8 GHz, and the number of acquired points is 6401. Then,
calibrated S21 of the high-temperature characterization setup described in the previous section is
recorded and shown in Fig. 4.11.
44
The measured fr and Qu of the SiCN resonator are plotted against temperature in Fig. 4.12. fr
decreases from 12.643 to 12.357 GHz as the temperature increases from 25 to 500oC. Therefore,
the average measurement sensitivity is found to be 0.547 MHz/oC. It should be noted that the
thermal expansion coefficient (CTE) of SiCN is very small, i.e. 5×10-6
/K; therefore its effect on
the fr is insignificant. The maximum Qu of 207.5 was measured at room temperature. Qu is
reduced to 45.1 at 500oC due to the increased losses within SiCN and metal. Ten independent
measurements are performed to verify the measurement repeatability. The mean value, standard
deviation and experimental uncertainty for both fr and Qu , and extracted εr and tanδ are listed in
Table II. It shows the standard deviation for fr is very small, no more than 0.02%, and for Qu, less
than 3%.
Figure 4.11: Measured S21 versus frequency at different temperatures.
12 12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8-45
-40
-35
-30
-25
-20
-15
Frequency (GHz)
S21 (
dB
)
25oC
100oC
200oC
300oC
400oC
500oC
25oC
500oC
45
Figure 4.12: Measured fr and Qu versus temperature.
Figure 4.13: Extracted εr and tanδ versus temperature.
2550 100 150 200 250 300 350 400 450 50012.35
12.40
12.45
12.50
12.55
12.60
12.65 f
r
Qunloaded
Temperature (oC)
f r (G
Hz)
20
50
80
110
140
170
200
Qu
nlo
ad
ed
2550 100 150 200 250 300 350 400 450 5003.70
3.74
3.78
3.82
3.86
3.90
εr
tanδ
Temperature (oC)
εr
0.000
0.005
0.010
0.015
0.020
0.025
tanδ
46
Figure 4.14: Measured and simulated S21 results of the SiCN sample characterization at 25 and
500oC
The extracted εr and tanδ of SiCN are plotted in Fig. 4.13, by performing parametric sweeps of
these two variables in HFSS simulations. The comparison between the measured and simulated
S21 results for 25 and 500oC characterization is shown in Fig. 4.14. It should be noticed that the
agreement between the two curves constituted an excellent confirmation of this method. The
dielectric constant and loss tangent of SiCN increase from 3.707 to 3.883 and from 0.0038 to
0.0213, respectively, within 25-500oC. It should be noted that these two values at the room
temperature agree well with the results obtained from the waveguide cavity method. In addition,
due to the lack of experimental data on the temperature-dependent metal conductivity, it is
assumed to be constant across the temperature range in HFSS simulations, which typically
12.0 12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8-45
-40
-35
-30
-25
-20
-15
-10
500oC
Measurement
SimulationS
21 (
dB
)
Frequency (GHz)
25oC
47
decreases versus temperature. Therefore, the actual loss tangent of SiCN at elevated temperatures
should be lower than the extracted values.
Figure 4.15:Measured fr versus temperature at the center temperature of (a) 100oC, (b) 200
oC, (c)
300oC, (d) 400
oC, and (e) 500
oC.
In order to quantify the temperature measurement accuracy using this method, the measurement
procedure is repeated in the range of ±20oC at center temperatures of 100, 200, 300, 400, and
500oC, with 5
oC increments. The measured fr versus temperature is shown in Fig. 4.15. To
compare the slope of these curves, all the plots in Fig. 4.15 use the same scale in the vertical axis.
It is observed the change of fr is progressively faster when temperature increases, which indicates
(a) (b) (c)
80 85 90 95 100 105 110 115 12012.590
12.599
12.608
12.617
12.626
12.635
12.644
f r (G
Hz)
Temperature (oC)
280 285 290 295 300 305 310 315 32012.516
12.525
12.534
12.543
12.552
12.561
12.570
f r (G
Hz)
Temperature (oC)
180 185 190 195 200 205 210 215 22012.548
12.557
12.566
12.575
12.584
12.593
12.602
f r (G
Hz)
Temperature (oC)
380 385 390 395 400 405 410 415 42012.460
12.469
12.478
12.487
12.496
12.505
12.514
f r (G
Hz)
Temperature (oC)
480 485 490 495 500 505 510 515 52012.353
12.362
12.371
12.380
12.389
12.398
12.407
f r (G
Hz)
Temperature (oC)
(d) (e)
48
higher temperature measurement sensitivity (resolution) at higher temperatures. It is noted that
Qu decreases versus temperature, which adversely affects the measurement sensitivity.
Table 4.1: Measured fr, Qu and extracted εr, tanδ with standard deviations (ζ) and uncertainty (su)
for SiCN sample at different temperatures
4.4 Characterization Up to 1000oC
At the temperature above 500oC the chromium oxide products, even though there is a 7-μm layer
of gold covering the chromium layer. Therefore, the CPW line with Cr-Au metallization cannot
work when the temperature is over 500oC. The platinum is more stable at high temperature and is
one of the best candidates for the metallization. Platinum (Pt) paste (5542 PRINT Grade, ESL
Electro Science) is used to metalize the CPW line and the Si4B1CN (sintered at 1000oC) sample.
Temperature(oC) fr (GHz) (ζ) (su) εr (ζ) (su) QU (ζ) (su) tanδ (ζ) (su)
25 12.6433 (0.02%)
(0.0007)
3.7071 (0.04%)
(0.0004)
207.5 (2.92%)
(1.9)
0.0038 (3.49%)
(0.0001)
100 12.5958 (0.01%)
(0.0005)
3.7357 (0.03%)
(0.0003)
169.7 (1.24%)
(0.6)
0.0048 (1.31%)
(0.0001)
200 12.5776 (0.01%)
(0.0002)
3.7466 (0.01%)
(0.0001)
144.6 (1.27%)
(0.6)
0.0063 (1.99%)
(0.0001)
300 12.5553 (0.01%)
(0.0002)
3.7601 (0.01%)
(0.0001)
100.4 (1.58%)
(0.6)
0.0093 (1.65%)
(0.0001)
400 12.4758 (0.01%)
(0.0004)
3.8081 (0.02%)
(0.0002)
73.1 (1.36%)
(0.2)
0.0130 (1.39%)
(0.0001)
500 12.3565 (0.02%)
(0.0007)
3.8826 (0.04%)
(0.0004)
45.1 (1.97%)
(0.2)
0.0213 (2.25%)
(0.0001)
49
The advantages of the metallization with Pt paste are relatively simple and low cost, compared
with evaporation and plating techniques.
Figure 4.16:Dimensions of (a) the Si4B1CN (1000oC) resonator and the coupling slot, (b) high
temperature connector and (c) holder. D = 9.37; H = 4.71; W1 = 3; H1 = 1; W = 1.5; L = 1.35; S =
0.4; L1 = 12.7; L2 = 8.64; D1 = 3.05; D2 = 1.32; D3 = 2.54; L3 = 15.24; L4 = 8.64; L5 = 4.57; L6 =
7.00; L7 = 10.92; H2 = 5.33; H3 = 8.64; H4 = 2.60; H5 = 1.00; H6 = 6.31. (unit: mm)
L1
L2
D1D2
L2L1
D3
H
D
H1W1
(a)
(b) (c)
L3
L3
L4
L5
L6
L7
H3
H2
H4
H5
D4
H6
50
Figure 4.17: (a) The furnace containing high-temperature measurement setup connected to the
high temperature cables. (b) Sample over CPW line outside the furnace. (c) Connector with
holder before soldering .(d) Sample and CPW line inside the furnace. (e) High temperature
connectors soldered onto the CPW line with the high temperature solder.
(a)
(b) (c)
(d) (e)
51
4.4.1 Metallization Process and Measurement Setup
The metallization process of the CPW line and the resonator is performed as follows. First, a
layer of Pt paste was painted over a piece of alumina board and the board was left at room
temperature for 10 minutes. Then, the board was dried at 100oC for 10 minutes, and the sample
was fired at 1000oC for 10 minutes in the MTI GSL-1600X furnace. The ramp rate for both
heating up and cooling down is 10oC/min. The thickness of formed Pt layer is approximately 40
μm. Slots with 0.4 mm in width were cut on the board using diamond saw to form the CPW line.
The dimension of CPW line with Pt metallization is the same as the one in Section III-B. The
procedure of the metallization of the resonator is the same as that for CPW line, except that the
opening slots on the sidewall of the sample were covered by the tape before being painted. The
dimensions of the resonator with coupling slots are shown in Fig. 4.16(a).
Figure 4.18:TRL calibration standards for the measurement
In order to ensure the measurement setup working above 1000oC, high temperature solder
(91Pb4Sn4Ag1In, melting point: 313oC, Indium Corporation), high temperature connectors
(600oC) and cables (1000
oC, Time Microwave Systems) are used. A copper holder was custom-
Thr
Open
Line
52
made for better connection between the CPW line and connector. The dimensions of the
connector and holder are shown in Fig. 4.16(b) and (c). The whole setup with close-up figures is
shown in Fig. 4.17. High temperature cables are also used as the extension of CPW line to
connect to the VNA, so the temperature at the ports of VNA is the room temperature.
The TRL calibration was conducted at each temperature before the measurements. The
dimensions of the TRL calibration standards shown in Fig. 4.18 is the same as the one in the
characterization up to 500oC . The temperature of the inside of furnace was increased from 25 to
1000oC during the measurement at the step of 100
oC when the temperature was above 100
oC.
Ten independent measurements were performed for each temperature.
4.4.2 Measurement Results
The measurement results of S21 for the Si4B1CN (1000oC) sample at different temperatures are
plotted in Fig. 4.19. The fr and Qu versus temperature extracted from the results of S21 are plotted
in Fig. 4.20. The fr decreases from 11.206 to 10.856 GHz and Qun decreases from 224.1 to 47.7,
when the temperature increases from 25 to 1000oC. Through HFSS simulations the εr and tanδ
are extracted and plotted in Fig. 4.21. The εr and tanδ of SiCN increase from 4.8169 to 5.1322
and from 0.0020 to 0.0186, respectively, within 25-1000oC. The type A experimental
uncertainties for εr and tanδ are not more than 0.0004 and 0.0001, respectively. The mean value,
standard deviation and experimental uncertainty for measured fr and Qu ,and extracted εr and
tanδ are listed in Table 4.2. In comparing the measured and simulated S21 results for 25 and
1000oC shown in Fig. 4.22, it is found that the consistent agreement between the two curves
confirms this characterization method.
53
Figure 4.19:Measured S21 versus frequency at different temperatures.
Figure 4.20:Measured fr and Qu versus temperature.
25 100 200 300 400 500 600 700 800 900 100010.8
10.9
11.0
11.1
11.2
11.3 f
r
Qunloaded
Temperature (oC)
f r (G
Hz)
20
50
80
110
140
170
200
230
260
Qu
nlo
ad
ed
10.6 10.7 10.8 10.9 11.0 11.1 11.2 11.3 11.4-55
-45
-35
-25
-15
-5
1000oC
S2
1 (
dB
)
Frequency (GHz)
25oC
54
Figure 4.21: Extracted εr and tanδ versus temperature.
Figure 4.22: Measured and simulated S21 results of the Si4B1CN sample characterization at
1000oC.
25 100 200 300 400 500 600 700 800 900 10004.7
4.8
4.9
5.0
5.1
5.2
εr
tanδ
Temperature (oC)
εr
0.000
0.004
0.008
0.012
0.016
0.020
0.024
tanδ
10.6 10.7 10.8 10.9 11.0 11.1 11.2 11.3 11.4-50
-40
-30
-20
-10
1000oC
Measurement
Simulation
S2
1 (
dB
)
Frequency (GHz)
25oC
55
Table 4.2: Measured fr, Qu and extracted εr, tanδ with standard deviations (ζ) and uncertainty (su)
for Si4B1CN (1000oC) sample at different temperatures
Temperature
(oC)
fr (GHz) (ζ) (su) εr (ζ) (su) QU (ζ) (su) tanδ (ζ) (su)
25 11.2065 (0.01%)
(0.0005)
4.8169 (0.03%)
(0.0004)
224.1 (1.71%)
(2.2)
0.0020 (5.80%)
(0.0001)
100 11.1880 (0.01%)
(0.0001)
4.8330 (0.01%)
(0.0001)
200.3 (1.20%)
(0.6)
0.0028 (2.68%)
(0.0001)
200 11.1553 (0.01%)
(0.0002)
4.8615 (0.01%)
(0.0002)
159.7 (1.05%)
(0.3)
0.0040 (2.34%)
(0.0001)
300 11.1275 (0.01%)
(0.0002)
4.8857 (0.01%)
(0.0002)
142.0 (0.48%)
(0.2)
0.0050 (0.76%)
(0.0001)
400 11.0871 (0.01%)
(0.0003)
4.9208 (0.02%)
(0.0003)
126.5 (0.80%)
(0.3)
0.0059 (1.06%)
(0.0001)
500 11.0557 (0.01%)
(0.0002)
4.9494 (0.01%)
(0.0002)
105.1 (0.30%)
(0.1)
0.0073 (0.29%)
(0.0001)
600 11.0343 (0.01%)
(0.0001)
4.9685 (0.01%)
(0.0001)
97.7 (0.76%)
(0.2)
0.0075 (1.46%)
(0.0001)
700 11.0118 (0.01%)
(0.0002)
4.9886 (0.01%)
(0.0002)
74.5 (1.77%)
(0.4)
0.0109 (1.45%)
(0.0001)
800 10.9095 (0.02%)
(0.0005)
5.0643 (0.03%)
(0.0004)
62.2 (0.99%)
(0.2)
0.0130 (1.40%)
(0.0001)
900 10.8925 (0.01%)
(0.0004)
5.0986 (0.02%)
(0.0004)
56.4 (0.72%)
(0.1)
0.0152 (0.94%)
(0.0001)
1000 10.8563 (0.01%)
(0.0004)
5.1322 (0.02%)
(0.0003)
47.7 (0.96%)
(0.1)
0.0186 (0.94%)
(0.0001)
56
4.5 Characterization Up to 1300oC
Due to the higher loss tangent of the ceramic materials at the higher temperature, the resonant
peak would fall below the noise level at the higher temperature. Therefore, it is harder to observe
the resonant peaks of the resonator by the weak coupling through the openings on its sidewall. In
order to keep the resonant peaks above the noise level, the slots under the resonator are cut for
stronger coupling. The dimensions of coupling slots for characterizing Si4B1CN (sintered at
1200oC) are shown in Fig. 4.23.
Through calibration was conducted at room temperature before the measurements. The
temperature of the inside of furnace was increased from 25 to 1300oC during the measurement at
the step of 50oC when the temperature was above 50
oC. The measurement results of S21 for
different temperatures are plotted in Fig. 4.24. The fr and Qu versus temperature extracted from
the results of S21 are plotted in Fig. 4.25 (a) and (b). The fr decreases from 11.2 to 10.6 GHz and
Qun decreases from 85 to 19, when the temperature increases from 25 to 1300oC. At 1300
oC the
resonant peak is closed to the noise floor and the Qun reaches the lowest value of 19. The
extracted εr and tanδ versus temperature are plotted in the Fig. 4.26. The εr and tanδ of SiCN
increase from 4.7561 to 5.3281 and from 0.0081 to 0.0434, respectively, within 25-1300oC.
57
Figure 4.23: (a) Schematic of the resonator over CPW line (b) dimensions of the coupling slots
(c) photos of sidewall and bottom with coupling slots for sample Si4B1CN (1200oC). (D=9.37;
H=4.71; W1=3; H1=1; W=1.5; L=1.35; S=0.4) (unit: mm)
Coupling Slot
H
D
H1W1
L
WS
(a)
(b)
(c)
58
Figure 4.24:S21 for temperature from 25oC to 1300
oC.
25oC
1300oC
10.5 10.6 10.7 10.8 10.9 11.0 11.1 11.2 11.3 11.410.4 11.5
-40
-30
-20
-10
-50
-5
Frequency(GHz)
S2
1(d
B)
59
Figure 4.25:(a) Extracted fr and (b) extracted Qun versus temperature.
25 100 200 300 400 500 600 700 800 900 100011001200130010.5
10.6
10.7
10.8
10.9
11.0
11.1
11.2
11.3f r(G
Hz)
Temperature (oC)
25 100 200 300 400 500 600 700 800 900 100011001200130010
30
50
70
90
110
Qu
n
Temperature (oC)
(a)
(b)
60
Figure 4.26:(a) Extracted εr and (b) extracted tanδ versus temperature .
25 100 200 300 400 500 600 700 800 900 10001100120013004.7
4.8
4.9
5.0
5.1
5.2
5.3
5.4
εr
Temperature (oC)
25 100 200 300 400 500 600 700 800 900 10001100120013000.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.045
tanδ
Temperature (oC)
(a)
(b)
61
4.6 Conclusion
The dielectric constant and loss tangent of high-temperature-stable Si(B)CN ceramic materials
have been characterized at microwave frequencies up to 1300oC. Two measurement setups for
500oC and 1000
oC characterization with TRL calibrations were described in the paper. More
than 10 measurements were done for all the characterizations to correctly estimate the statistical
quantities: mean value and Type A experimental uncertainty. The accurate characterization of
SiCN materials is critical for developing wireless passive sensors in harsh environment
applications. Wireless passive sensors using SiCN materials will be developed for combustion
turbines in the next chapter. Some sections inside turbines used in the power generation are at the
high temperatures (from a few hundred to 1000oC and above), so these proposed sensors can be
used for temperature sensing.
62
CHAPTER 5 HIGH TEMPERATURE MATERIALS APPLICATIONS
5.1 Introduction
Wireless passive sensing systems are very attractive in a wide range of applications. The
advantages of wireless passive sensing include battery-free operation, higher reliability, low
maintenance cost, and a better chance of surviving harsh environments. The temperature
dependency of the dielectric constant of SiCN materials is demonstrated at both low frequencies
and microwave frequencies in the previous chapters. Based on this temperature-dependent
dielectric constant behavior, three different types of sensors, corresponding to different antenna
configurations, are designed and the prototypes are fabricated and tested. All of the sensors
successfully perform at temperatures over 1000oC.
5.2 Two-antenna Sensor
5.2.1 Wireless Sensing System Using Passive Microwave Resonators
A wireless sensing technique based on passive microwave resonators is proposed. The resonant
frequency of the microwave resonator is related to temperature or gas properties. By measuring
the resonant frequency of the resonator using a wireless technique, the information regarding
temperature or gas properties can therefore be retrieved. To prove this concept, a wireless
sensing system is designed and measured using a spiral resonator. However, this developed
technique is general enough to be implemented to high-temperature or gas sensing applications
by using resonators made of either high-temperature ceramics or gas sensitive materials [43].
63
Figure 5.1:(a) Illustration of a wireless passive sensor system. (b) Equivalent circuit of the sensor
system in (a).
Fig. 5.1(a) illustrates the proposed wireless sensing system based on a passive microwave
resonator. The passive sensor is composed of two antennas and a microwave resonator. This
sensor system sends out a signal from Antenna 1 to Antenna 3. Since there is a resonator in
between Antennas 3 and 4, only the energy at the resonant frequency passes through this
resonator and reaches Antenna 2 via Antenna 4, representing a peak in S21 measured by a vector
network analyzer.
This wireless sensing system is modeled by an equivalent circuit as shown in Fig. 5.1(b). The
microwave resonator is represented by a parallel RLC network. The coupling between the
resonator and the transmission line can be characterized using a transformer. Since there is free-
space path loss due to the transmission of energy between the antennas, we can use a lossy
transmission line to model this phenomenon. Based on this equivalent circuit model, it is
expected that S21 has a peak in the frequency domain which corresponds to the resonant
(a) (b)
Passive
Sensor
64
frequency of the microwave resonator, though the magnitude of S21 is reduced due to the free
space path loss.
Figure 5.2: Wideband antenna geometry. (G=48mm, L=28mm,W=24.2mm,S=0.15mm,h=14mm,
g=0.2mm, Wf=3mm)
Figure 5.3:4.02 GHz resonator geometry. (W1=5.3mm, W2=3mm, W3=0.3mm, W4=0.5mm,
g1=0.5mm, g2=1.6mm, g3=0.2mm, g4=0.15mm, g5=1.95mm)
(a) (b)
65
In real sensing applications, the resonant frequency of the resonator changes versus temperature
or gas properties. Therefore, the antenna used in this system should cover the range of possible
resonant frequencies. The antenna structure published in [44] is used here to achieve a 35%
factional bandwidth with a center frequency of 3.82 GHz. A spiral resonator structure [45] is
chosen for compact size and high quality factors and designed to achieve resonant frequencies of
3.59 GHz and 4.02 GHz, respectively. Both antennas and resonators are designed on 60-mil-
thick Rogers RO3003 substrates. The geometries of the antenna and spiral resonator are
presented in Fig. 5.2 and 5.3 in detail.
This wireless sensing system is set up as shown in Fig. 5.1(a) inside an anechoic chamber. The
distance between Antennas 1 and 2 is set to be 28 cm. In order to study the sensing range of this
wireless system, the spacing between antennas 1 and 3, which is same as that between antennas 2
and 4, is swept from 3 cm to 10 cm. The measured S21 of the system with resonators at 3.59 and
4.02 GHz are plotted in Fig. 5.4 and 5.5, respectively. To verify that the measured resonant
frequencies of the wireless system are same as those of the resonators themselves, S21 for each
resonator is also plotted for comparison. Very close agreement (within 0.6%) has been found
among measured resonant frequencies [43].
66
Figure 5.4:Measured S21 for the 3.59 GHz resonator only and the wireless system with antenna
spacing of 3 cm, 6cm, 8 cm.
Figure 5.5:Measured S21 for the 4.02 GHz resonator only and the wireless system with antenna
spacing of 3 cm, 6cm, 8 cm.
67
Figure 5.6:Direct transmission between Antennas 1 and 2 for Co-Pol and X-Pol.
Figure 5.7:Comparison of wireless sensing between Co-Pol and X-Pol cases with an antenna
spacing of 4 cm.
68
In the demonstration shown earlier, antennas 1 and 3 are vertically polarized while antennas 2
and 4 are horizontally polarized. The direct energy coupling between antennas 1 and 2 has an
adverse effect on the measured results. Fig. 5.6 shows the transmission between Antennas 1 and
2 for both co-polarization and cross-polarization cases. The cross-polarization case gives
approximately 10-dB more isolation than the co-polarization case. This phenomenon is further
proved by observing a case in which the spacing between antennas 1 and 3 is 4 cm. As shown in
Fig. 5.7, the resonant peak is unidentifiable for the co-polarization case, but clearly visible for
the cross-polarization case [43].
A wireless sensing system based on a microwave resonator has been successfully demonstrated.
The sensed resonant frequency is identical to the resonant frequency of the resonators. A sensing
range up to 8 cm between antenna pairs was achieved. It was found that the placement of antenna
orientations has a noticeable effect on the sensing range. This wireless sensing system will be
used to sense very high temperatures using ceramic-based microwave resonators [43].
5.2.2 Sensor Design
The wireless passive detection method based on microwave resonators was presented in the last
section. Two antennas were connected to a planar resonator to enable wireless detection of the
resonant frequency. However, there were three issues associated with that approach for the
aforementioned applications. First, the antennas used in that approach were bidirectional, causing
undesired radiations and loss of energy. Second, the resonator was made of non-high-
69
temperature-stable dielectric materials. Third, due to the planar geometry of the resonator, the Q
factor of the resonator was very limited, which restricted the sensing resolution [46].
Figure 5.8:Flowchart showing the wireless sensing mechanism.
To overcome the aforementioned limitations, a new sensing mechanism is proposed here using
two pairs of antennas to wirelessly measure the resonant frequency of a dielectrically-loaded
cavity resonator as shown in Fig. 5.8. The cavity resonator is fabricated using SiCN ceramic
material and coupled to CPW lines through two coupling slots. Two pairs of antennas are
implemented at each side of the resonator. Within each antenna pair, one is connected to the
CPW line and the other is connected to one port of the network analyzer. A wide-band signal is
transmitted through the antenna pairs and passes the resonator via the coupling slots. Maximum
Network Analyzer
Signal sent
out by VNASignal received
by VNA
Resonator(f0)
Antenna CPW line SMA ConnectorCoaxial line
70
transfer of energy occurs at the resonant frequency of the resonator and is observed as a peak in
the S21. This wireless detection of the resonant frequency will be used to form a temperature
sensor due to the temperature-dependent dielectric constant of SiCN ceramic materials employed
in the resonator [46].
Figure 5.9:(a) Passive temperature sensor with a SiCN cavity resonator coupled to CPW lines. (b)
Top and (c) side view of the sensor. S = 1.5; G = 0.4; W = 10; T = 0.635; L = 149.4; D = 9.41; H
= 5.01; H1 = 1; W1 = 3 (unit: mm) (d) Patch antenna. Wa = 9.2; La = 7.7; Lm = 2.6; Ws = 0.4; Wf=
2.45.(unit: mm)
Details of the proposed high-temperature sensor based on a SiCN resonator are presented as
follows. As seen in Fig. 5.9, a SiCN cavity resonator is placed over CPW lines. The substrate
material for the CPW line is alumina. All the dimensions are listed in Fig. 5.9. Two coupling
W1
H1
H
T
L
GSWD
(a) (b)
(c) (d)
La
Lm
Wa
Ws
Wf
71
slots on the sidewall of the resonator are created during the metallization procedure of the SiCN
resonator, which allows for the magnetic coupling between the short-ended CPW lines and
resonator. For the dielectric cavity resonators studied herein, H/D < 1.015, therefore the
dominant resonant mode for this resonator is TM010. SMA connectors are soldered to the CPW
lines for connecting the antennas.
The resonant frequency of the SiCN resonator is determined by the resonator size and the
dielectric constant of SiCN. When the dielectric constant of SiCN changes versus temperature,
the resonant frequency of the resonator changes accordingly.
Figure 5.10:Simulated S11 of the patch antenan shown in Fig. 5.9(d) and S21 of the structure
shown in Fig. 5.9(a).
A microstrip patch antenna is designed for the sensor. Fig. 5.9(d) shows the geometry of the
proposed patch antenna. The antenna is fabricated on a Rogers RT/Duroid 5880 substrate with a
thickness of 0.787 mm. The center frequency and bandwidth of the antenna is 12.6 GHz and
11.5 11.7 11.9 12.1 12.3 12.5 12.7 12.9 13.1 13.3 13.5-50
-40
-30
-20
-10
0
S1
1(d
B)
Antenna
Resonator
Frequency (GHz)
S2
1(d
B)
-50
-40
-30
-20
-10
0
72
approximately 3%, respectively. The gain of this antenna is more than 8 dB over the bandwidth.
The simulated antenna return loss is shown in Fig. 5.10. The transmission of the structure shown
in Fig. 5.9(a) is also plotted in Fig. 5.10. It is apparent that the resonant frequency of the sensor is
within the antenna bandwidth.
Figure 5.11:Entire wireless sensing system.
The wireless sensing system is illustrated in Fig. 5.11. The passive sensor is composed of two
antennas and a SiCN cavity resonator. This sensor receives a wide-band signal from Antenna 3,
which is coupled to Antenna 1. Since there is a resonator between Antennas 3 and 4, only the
energy at the resonant frequency passes through this resonator and then reaches Antenna 2 via
Antenna 4, representing a peak in S21, as shown in Fig. 6. The distance between antennas 1 and 3
is kept the same as that between antennas 2 and 4. Although the resonator is relatively weakly
coupled to the CPW lines, a peak in the transmission can still be clearly observed due to the high
Q factor of the resonator [46].
73
Figure 5.12:Simulation result for the sensing system at the distance of 20mm between antennas
in Fig. 5.11.
5.2.3 Fabrication and Measurement up to 500oC
Details about the fabrication process and measurement results are provided in this section. Fig.
5.13(a) shows the fabricated cavity resonator over the CPW line. In order to form a cavity
resonator made of the SiCN ceramic material, metallization of the resonator is performed in two
steps. First, a 0.3-μm-thick Cr layer is evaporated on the sample. Five evaporations from
different directions are necessary to cover the entire surface of the sample. Then Au
electroplating is performed to create a 7-μm-thick metal layer in order to minimize the conductor
loss. During the metallization steps, the coupling slots are covered by 3 by 1 mm tapes. The
CPW lines over the alumina substrate are created using the same metallization procedure and
patterned using photolithography. The height and diameter of the SiCN cavity resonator is 5.01
11.5 11.7 11.9 12.1 12.3 12.5 12.7 12.9 13.1 13.3 13.5-60
-55
-50
-45
-40
-35
-30
-25
-20
Frequency (GHz)
S2
1 (
dB
)
74
and 9.42 mm, respectively. It should be noted that Cr/Au is used here to demonstrate the
proposed wireless sensing concept. For turbine applications, Pt will be used in order for the
sensors to work above 1000oC. making it a promising candidate for future high-temperature
sensor development. The fabricated antenna is shown in Fig. 5.13(b). Four same antennas are
fabricated for this research.
Figure 5.13:(a) Fabricated cavity resonator over the CPW line (b) fabricated patch antenna; all
four antennas are the same
The structure shown in Fig. 5.9(a) is measured and compared with simulations as shown in Fig.
5.14. The reflection coefficient of the fabricated antenna is also measured and shown in the same
figure. Good agreement between the simulation and measurement has been observed.
After demonstrating the performance of resonator and antennas separately, the final wireless
sensing setup is formed as shown in Fig. 5.15. Each end of the CPW lines is connected to an
antenna. Those two antennas are coupled to two other antennas which are connected to the VNA.
Fig. 5.16 compares the measured S21 for the cases with or without antennas. It is observed that
(a) (b)
75
for both cases the maximum transfer of energy happens at the resonant frequency of the
resonator, i.e. 12.6 GHz. It shows that we can successfully measure the resonant frequency
wirelessly at the cost of about 11-dB loss. The same measurement is performed for various
distances between the antenna pairs. As shown in Fig. 5.17, as the distance increases, the level of
received signal decreases. Maximum sensing distance around 40 mm is achieved with this
measurement setup. It should be noted that the measured loaded Q reduces from 170 down to
150 when the distance is increased from 10 up to 40 mm.
Figure 5.14:Measured and simulated S11 of the patch antenna shown in Fig. 9(d) and S21 of the
structure shown in Fig. 9(a).
11.5 11.7 11.9 12.1 12.3 12.5 12.7 12.9 13.1 13.3 13.5-50
-40
-30
-20
-10
0
Frequency (GHz)
S1
1(d
B)
Antenna_measurement
Antenna_simulation
Resonator_simulation
Resonator_measurement
-50
-40
-30
-20
-10
0
S2
1(d
B)
76
Figure 5.15: Final wireless passive sensing setup.
Figure 5.16: Measured S21 with or without the antennas.
11.5 11.7 11.9 12.1 12.3 12.5 12.7 12.9 13.1 13.3 13.5-50
-45
-40
-35
-30
-25
-20
-15
-10
Frequency(GHz)
S2
1(d
B)
Without antenna
Antenna@20mm
77
Figure 5.17: Measured S21 for sensing distances from 10 to 45 mm.
Figure 5.18: Measured S21 for different temperatures.
11.5 11.7 11.9 12.1 12.3 12.5 12.7 12.9 13.1 13.3 13.5-60
-55
-50
-45
-40
-35
-30
-25
-20
Frequency (GHz)
S2
1(d
B)
d=10mm
d=15mm
d=20mm
d=25mm
d=30mm
d=35mm
d=40mm
d=45mm
d=10mm
d=45mm
12.3 12.4 12.5 12.6 12.7 12.8-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
S2
1 (
dB
)
Frequency (GHz)
25oC
50oC
100oC
150oC
200oC
250oC
300oC
350oC
400oC
450oC
500oC
78
By enhancing the coupling between the resonator and CPW lines, the wireless sensing range can
be further increased. To demonstrate high-temperature sensing, the CPW lines and resonator is
placed inside a furnace with controllable temperature. The two ends of the CPW lines are outside
the furnace at a much lower temperature, allowing for soldering SMA connectors and connecting
antennas. Through calibration was done at room temperature. The results of S21 for different
temperatures are recorded, as shown in Fig.5.18. It is observed that the resonant peak moves to
lower frequency for higher temperature, which is the same as we estimated before.
5.2.4 Fabrication and Measurement up to 1000oC
Higher temperature measurement up to 1000oC is realized using Si4B1CN (1000
oC) sample. The
measurement setup and resonator dimensions are shown in Fig. 5.19. The wider bandwidth
antennas are used for covering the variation range of resonator because of the larger range of the
measurement temperature. The detail of the antenna, such as dimensions will be shown in the
next section. In this setup, the high temperature cable is used to connect the antennas and CPW
line in order to keep the antenna at low temperature. Through calibration is done at room
temperature before the measurement. The distance between a pair of face-to-face antennas was
kept at 10 mm during the measurement. The temperature of the inside of furnace was increased
from 25 to 1000oC at the step of 50
oC when the temperature was over 50
oC. The measured S21
results are shown in Fig. 5.20 (a). The measurements were repeated twice. The extracted fr, and
Qu, versus temperature are plotted in Fig. 5.20 (b) and (c).
79
Figure 5.19: (a) Passive temperature sensor with a Si4B1CN (1000oC) cavity resonator coupled to
CPW lines, (b) side view of the sensor, and (c) wireless passive sensing setup for room
temperature. T = 0.635; D = 9.37; H = 4.71; H1 = 1; W1 = 3 (unit: mm).
Resonator
CPW line
Antenna
Connector
1
2
3
4
D
H
TW1
H1
(a) (b)
(c)
80
Figure 5.20: (a) Measured S21 for different temperatures, extracted (b) fr and (c) Qun versus
temperature.
25 100 200 300 400 500 600 700 800 900 10000
30
60
90
120
150
180
210
240
Qu
n
Temperature (oC)
Meas1
Meas2
25 100 200 300 400 500 600 700 800 900 100010.85
10.90
10.95
11.00
11.05
11.10
11.15
11.20
11.25
f r(GH
z)
Temperature (oC)
Meas1
Meas2
(a)
(b)
25oC
1000oC
10.7 10.8 10.9 11.0 11.1 11.2 11.3 11.4 11.510.6 11.6
-40
-30
-20
-50
-10
Frequency (GHz)
S2
1 (
dB
)
(c)
81
5.3 One-antenna Sensor
5.3.1 Sensor Design
In this section, a cavity resonator connected with a patch antenna is proposed to work as a
wireless passive temperature sensor as shown in Fig. 5.21. The resonator is a cylindrical ceramic
sample coated with a metal layer. The ceramic materials studied here is SiBCN, which is SiCN
with Boron doping. The resonant frequency of this resonator sensor changes versus temperature,
which is mainly determined by the temperature-dependent dielectric properties of the ceramics.
This resonant frequency can be wireless detected by another patch antenna through the patch
antenna connected to the resonator [47].
Figure 5.21: Schematic of the wireless temperature sensor. D=9.2; H=4.7; W1=3; H1=1 (unit:
mm).
82
The key part of the temperature sensor is the cylindrical cavity resonator loaded with high
temperature ceramics with the temperature dependent dielectric property. The resonator is
metalized with Pt paste with the conductivity of 2×106 S/m. Due to the fabrication limitations of
the dielectric material samples, the cylinder diameter and thickness are fixed at D = 9.2 mm and
H = 4.7 mm, respectively. The dominant mode for this resonator is TM010 mode. The dielectric
constant and loss tangent are 4.8 and 0.003 respectively, which are characterized at room
temperature. The resonance frequency and unloaded Q factor of the closed cavity resonator can
be calculated to be 11.38 GHz and 236 based on analytical equations. The slots are opened on
the sidewall of the resonator for the magnetic field coupling with coplanar waveguide (CPW)
line. The dimensions of slots are shown in Fig. 5.21. The CPW line is fabricated using alumina
board metalized with a layer of Pt paste. The center conductor and slot of the CPW line are 1.5
and 0.4 width in mm in order to keep the characteristic impedance is nominally 50 Ω.
A microstrip patch antennas is designed for receiving and sending signals in order to make the
sensor working wirelessly. The center frequency of the antenna is 11 GHz and the bandwidth is
approximately 9%, which can cover the variation range of the resonant frequency of the
resonator due to the change of the ambient temperature. The gain of this antenna is more than 7
dB over the bandwidth.
With the temperature increasing, the dielectric constant of SiBCN will monotonically increase.
Therefore, the resonant frequency of the resonator decreases against temperature. The
relationship of the resonant frequency versus temperature is the basis of the sensing mechanism
in this paper. The study of the resonant frequency versus dielectric constant is done based on
83
HFSS simulation. In the simulation, the dielectric constant of SiBCN ceramic swept from 4.8 to
5.1, while the distance between antennas is kept as 30mm. In order to observe the resonance
frequency of the resonator clearly, time-domain (TD) gating is applied to S11 results to identify
the resonant peaks. The S11 results after TD gating corresponding to different dielectric constants
are plotted in Fig.5.22. The peaks of S11 represent the resonant frequencies corresponding to
different dielectric constants. The resonant frequency versus dielectric constant is extracted and
plotted in Fig. 5.23. It can be noticed that the resonant frequency goes down when the dielectric
constant increases [47].
Figure 5.22: Simulated S11 of the patch antenna for successive distances away from the sensor
after TD gating.
84
Figure 5.23: Simulated resonant frequency decreases with the increase of dielectric constant of
the SiBCN material.
5.3.2 Fabrication and Measurement up to 1000oC
A microstrip patch antenna is designed and fabricated for the sensor. Fig. 5.24 shows the
geometry of the proposed patch antenna. The antenna is fabricated on a Rogers RT/Duroid 5880
substrate with a thickness of 62 mils. The center frequency of the antenna is 11 GH and
bandwidth is approximately 9%, which covers the change range of the SiBCN resonator. The
gain of this antenna is more than 7 dB over the bandwidth.
85
In order to form a waveguide cavity resonator, the opening slots (3×1mm) on the sidewall of the
sample were covered by the tape, and then , a layer of platinum paste (ESL 5542) is applied on
all the faces of the cylindrical samples. Then the samples covered by platinum paste are dried at
100ºC for 10 minutes and sintered at 1000ºC for 10 minutes. The ramp rate during the sintering
is 10ºC/min. The formed platinum layer is approximately more than 30 µm thick.
Figure 5.24:Dimensions of the patch antenna. (a) Top view, (b) Side view, and (c) Fabricated
antenna. L=8.3;W=10.8;D1=1.27;D2=4.32;D3=6.5;T1=1.57 (unit: mm).
In order to ensure the measurement setup working above 1000oC, high temperature solder
(91Pb4Sn4Ag1In, melting point: 313oC, Indium Corporation), high temperature connectors
(600oC) are used. A copper holder was custom-made for better connection between the CPW line
and connector. The sensor is shown in Fig.5.25 (a). A high temperature furnace is used to
control the temperature of the sensor from 25 to 1000oC as shown in Fig. 5.25 (b). A K-type
thermocouple (Omega HH11) show in Fig. 5.25 (c) is used to monitor the temperature inside the
heat pad and provide the feedback to the temperature controller. Antenna patch antenna is used
for the wireless sensing. One port Short-Open-Load (SOL) calibration is performed at the end of
D2
D1
D3
W
L
D3
T1
(a) (b) (c)
86
the coaxial cable and proper TD gating is applied using an Agilent 40-GHz PNA-L (N5230A).
Thus, the real-time resonant frequency of the sensor can be measured wirelessly for different
temperatures.
The measured S11 responses of the Si4B1CN (1000oC) temperature sensor at different
temperatures is plotted in Fig. 5.26. Each curve corresponds to a single temperature in the 25-
1000oC range. At each temperature, there is a unique peak with the highest signal level. This
peak corresponds to the resonant frequency of the sensor. The resonant frequencies and Qs of the
Si4B1CN (1000oC) sensor for temperatures from 25 to 1000ºC are plotted in Fig. 5.27. The
resonant frequency decreases from 11.44 to 10.99 GHz in this temperature range, which
corresponds to a measured sensor sensitivity of 0.46 MHz/ºC. The Q decreases from 194 at room
temperature to 39 at 1000oC. As the temperature increases, Q is reduced due to the increasing
loss of both ceramic and platinum. The distances between two antennas is 10 mm. The sensing
distance can be increased by increasing the output power from the transmitter.
87
Figure 5.25: (a) Sensor outside the furnace, (b) Measurement setup with furnace, and (c)
Resonator and CPW line inside the furnace.
(a)
(b)
(c)
88
Figure 5.26:Measured S11 responses of the sensor at different temperatures Si4B1CN (1000oC).
Figure 5.27: Measured resonant (a) frequency and (b) Q of the Si4B1CN (1000oC) sensor
10.6 10.8 11.0 11.2 11.4 11.6 11.810.4 12.0
-60
-50
-40
-30
-20
-70
-10
Frequency (GHz)
S11 a
fter
TD
gating (
dB
)25
oC
1000oC
25 100 200 300 400 500 600 700 800 900 100040
60
80
100
120
140
160
180
200
Q
Temperature (oC)
25 100 200 300 400 500 600 700 800 900 100010.9
11.0
11.1
11.2
11.3
11.4
11.5
f r (G
Hz)
Temperature (oC)
(a) (b)
89
5.4 Integrated Slot Antenna Sensor
5.4.1 Sensor Design
Figure 5.28:Schematic of the wireless temperature sensor.
In this section, a cavity resonator integrated with a slot antenna is proposed to work as a wireless
passive temperature sensor as shown in Fig. 5.28. This cylinder resonator is formed by coating a
dielectric substrate with a metal layer. The resonant frequency of this resonator sensor changes
versus temperature, which is primarily determined by the temperature-dependent material
properties of the substrate. This resonant frequency can be wireless detected by an open-ended
waveguide (OEWG) antenna through the integrated slot antenna. The seamless integration of slot
antenna and resonator, demonstrated in [48][49], is able to provide efficient energy coupling
between the sensor and free space without the need of additional volume for the antenna. It is
noted that this technique was also used to realize integrated filter/antennas with compact size and
90
high efficiency [50][51][52]. The sensing mechanism shown in Fig. 5.28 is suitable for
implementation over turbine blades which function as a large ground plane.
This temperature sensor is based on a cylindrical microwave cavity resonator. Due to the
fabrication limitations of the dielectric material for this sensor, the cylinder radius and thickness
are fixed at a = 4.5 mm and h = 4.5 mm, respectively. The dominant mode for this resonator is
TM010 mode and its resonant frequency is given by:
(5.1)
In which a is the radius of the cavity resonator and p01 = 2.405. By setting εr = 5.8 and tanδ =
0.01, the fundamental resonant frequency is 10.77 GHz and unloaded Q factor Qu is 86.6 from
eigenmode simulation in HFSS.
In order to construct the wireless sensor, a slot antenna is integrated with the aforementioned
cavity resonator as the Tx/Rx antenna. The simplified circuit model for the resonator/antenna is
presented in Fig.5.29. R0 shows the dissipation due to metallic and dielectric losses in the cavity
resonator, while Rr indicates the radiation loss from the slot antenna. L and C, from both the
resonator and the slot antenna, represent the elements for energy storage. To maximize energy
coupling for the resonator through the slot antenna, the radiation Q factor Qrad of the slot antenna
should be equal to the unloaded Q factor Qu of the resonator [53]:
Critical coupling requirement:
Qrad=Qu (5.2)
91
To design the antenna for the critical coupling requirement, a weak coupling excitation of the
cavity resonator is performed by two coaxial ports as shown in Fig. 5.30. These two ports affect
little on the resonant frequency and will not be included in the final sensor structure. The slot
antenna has the length of La and its position has a distance d away from the cylinder center. The
loaded Q factor QL is simulated by the two coaxial ports in HFSS. Then the external Q factor Qext
which is equal to the Qrad, can be calculated by
(5.3)
Figure 5.29:Circuit model for the integrated slot antenna/resonator.
To achieve the desired Qext value, slot position and length are considered as the variable
parameters. The design curve of Qext is plotted in Fig. 5.31 for different combinations of antenna
length La and position d. It can be observed that Qext decreases when the slot antenna becomes
longer or be further away from the center. This means more current distributed on the cavity top
LCR0Rrextu
extures
s
s
s+
s-
Free space Radiation Dissipation Resonator
92
surface is cut off by the slot antenna, resulting in stronger radiation coupling. Finally, La = 6.8
mm and d = 2 mm is selected to match Qu for maximum energy coupling.
Figure 5.30:Top view (a) and 3-D view (b) of the integrated slot antenna and cavity resonator.
The two coaxial ports of weak coupling are for the slot antenna design. w = 0.45 mm, r = 0.174
mm, R = 0.4 mm and L = 2 mm.
h
wLa
d
r R
L
Slot
antenna
Weak
coupling
coaxial
portsa
(a)
(b)
93
At this level, the design is finalized and the sensor’s dimensions are fixed. To detect the resonant
frequency of the sensor wirelessly, an X-band (8-12 GHz) open-ended waveguide (OEWG) is
used as the interrogator antenna, which exhibits a gain of 6 dBi around 10 GHz. Fig. 5.32 plots
simulated S11 responses for successive sensing distances between the OEWG and sensor. The
resonant behavior of the sensor is difficult to distinguish due to superimposition of reflections
from the OEWG, scatterings from the sensor structure, and the re-radiated waves from the
resonator. Therefore, it is necessary to apply time-domain (TD) gating to isolate the response
from the sensor itself.
Figure 5.31:Simulated external Q factor for different slot antenna positions d and length La.
6.5 6.6 6.7 6.8 6.9 7.0
70
80
90
100
110
Exte
rnal Q
La (mm)
d = 1.6 mm
d = 1.8 mm
d = 2.0 mm
94
Figure 5.32:Simulated S11 of the OEWG for successive distances away from the sensor before
TD gating.
Figure 5.33:Simulated TD responses of OEWG with and without the sensor for 20 mm distance.
9.0 9.5 10.0 10.5 11.0-16
-15
-14
-13
-12
-11
-10
20 mm
30 mm
40 mm
50 mm
S1
1 (
dB
)
Frequency (GHz)
0 2 4 6 8 10 12-120
-100
-80
-60
-40
-20
S11 (
dB
)
Time (ns)
with the sensor
without the sensor
Gating window
95
The S11 responses of the OEWG with and without the sensor are compared in time domain in Fig.
5.33 for a sensing distance of 20 mm. The main peak right after 0 ns is due to the reflection at the
open aperture of the OEWG. For the case without the sensor, S11 quickly drops to the noise floor.
When the sensor is present, it absorbs the incident wave from the OEWG and then re-radiates
back to the OEWG in a periodically decaying manner. A gating window should be defined to
isolate the response from the sensor. For a sensing distance of 20 mm, the gating start time is set
to be 0.7 ns. The gating stop time is set to be larger or equal to 11 ns where the two curves in Fig.
5.33 intersect each other. Similar gating procedures are carried out for different sensing distances
of 30, 40 and 50 mm.
Figure 5.34:Simulated S11 of the OEWG for successive distances away from the sensor after TD
gating.
Finally, TD-gated signals are transformed back into frequency domain and shown in Fig. 5.34.
The resonant frequency of the sensor can be clearly identified and is independent of the sensing
9.0 9.5 10.0 10.5 11.0
-60
-50
-40
-30
-20
S11 (
dB
)
Frequency (GHz)
20 mm
30 mm
40 mm
50 mm
96
distance. Compared with resonant frequency of 10.77 GHz for the resonator without the slot, the
resonant frequency for the sensor is reduced to 9.99 GHz by the loading effect from the slot. The
QL value of 44.6 is consistent with antenna design for critical coupling and the mismatch can be
neglected, even though the Qext and Qu are simulated at different frequencies. The maximum
sensing range between the interrogator antenna and the sensor is 80 mm by simulation.
Figure 5.35:Simulated resonant frequency decreases with the increase in dielectric constant of
the SiCN material.
When the temperature increases, the dielectric constant of SiCN will monotonically increase.
Therefore, the resonant frequency of the resonator decreases against temperature. This resonant
frequency and temperature relationship is the basis of the sensing mechanism for the wireless
passive sensors presented herein. The dielectric constant of SiCN ceramic is swept from 5.3 to
5.8 in HFSS simulations to prove the concept. The S11 responses of the OEWG after TD gating
9.0 9.5 10.0 10.5 11.0
-60
-50
-40
-30
S11 (
dB
)
Frequency (GHz)
r = 5.6
r = 5.8
r = 6.0
r = 6.2
97
are plotted in Fig.5.35 for different dielectric constants. The downshift tendency of resonant
frequency is shown in Fig.5.36.
Figure 5.36:Simulated resonant frequency decreases with the increase in dielectric constant of
the SiCN material for 30 mm distance
5.4.2 Fabrication and Measurement up to 1000oC
Two sensors are fabricated and measured in this study. One is made of Si6B1CN (sintered
1000oC) and another is made of Si4B1CN (sintered 1100
oC). In order to form a waveguide cavity
resonator, a layer of platinum paste (ESL 5542) is applied on all the faces of the cylindrical
samples. Then the samples covered by platinum paste are dried at 100ºC for 10 minutes and
sintered at 1000ºC for 10 minutes. The ramp rate during the sintering is 10ºC/min. The formed
platinum layer is approximately more than 30 µm thick. The slot antenna is formed on the top
5.5 5.6 5.7 5.8 5.9 6.0 6.1 6.2 6.39.7
9.8
9.9
10.0
10.1
10.2
R
esonant fr
equency (
GH
z)
Dielectric constant
98
surface of the cavity by milling process as shown in Fig. 5.37. The fabricated sensors with
dimensions are shown in Fig. 5.38.
Figure 5.37:Slot antenna fabrication by milling
Figure 5.38: (a)Top and (b) side views of the sensor and dimensions. (Si6B1CN (1000oC):
D=8.59, H=4.48, d=2.0, w=0.45, L=6.8; Si4B1CN (1100oC): D=9.24, H=4.82, d=2.0, w=0.45,
L=7.0) (unit: mm)
wslotL
d
DH
D
(a) (b)
99
A heat pad with the diameter of 2 inches (Micropyretics Heaters International Inc.) is used to
control the temperature of the sensor from 25 to 1050ºC/1150oC as shown in Fig. 5.39(a). A
piece of alumina board is placed over the heat pad to prevent air convection and stabilize the
temperature of the sensor as shown in Fig. 5.39(b). A K-type thermocouple (Omega HH11) show
in Fig. 5.39 (c) is used to monitor the temperature inside the heat pad and provide the feedback to
the temperature controller.
The X-band open-end waveguide (OEWG) is used for the wireless sensing. One port Short-
Open-Load (SOL) calibration is performed at the end of the coaxial cable and proper TD gating
is applied using an Agilent 40-GHz PNA-L (N5230A). Thus, the real-time resonant frequency of
the sensor can be measured wirelessly for different temperatures.
100
Figure 5.39: (a) Measurement set-up (b) The sensor is placed inside the heat pad with alumina
board cover. (c) Inside of the heater without alumina board cover.
(a) (b)
(c)
101
The measured S11 responses of the OEWG at different temperatures for Si6B1CN (sintered
1000oC) and Si4B1CN (1100
oC) temperature sensors are plotted in Fig. 5.40. Each curve
corresponds to a single temperature in the 25-1050/1150oC range. At each temperature, there is a
unique peak with the highest signal level. This peak corresponds to the resonant frequency of the
sensor. The resonant frequencies and Qs of the Si6B1CN (sintered 1000oC) sensor for
temperatures from 25 to 1050ºC are plotted in Fig. 5.41. The resonant frequency decreases from
11.05 to 10.71 GHz in this temperature range, which corresponds to a measured sensor
sensitivity of 0.33 MHz/ºC. The Q decreases from 82 at room temperature to 30 at 1050oC. For
the Si4B1CN (sintered 1100oC) sensor was measured from room temperature to 1150
oC. The
resonant frequencies and Qs of this sensor for this temperature range are plotted in Fig. 5.42. The
resonant frequency decreases from 10.39 to 10.06 GHz in this temperature range, which
corresponds to a measured sensor sensitivity of 0.30 MHz/ºC. The Q decreases from 52 at room
temperature to 25 at 1050oC. As the temperature increases, Q is reduced due to the increasing
loss of both ceramic and platinum. The distances between the OEWG and Si6B1CN (sintered
1000oC) and Si4B1CN (1100
oC) temperature sensors are 20 and 10 mm, respectively. The
sensing distance can be increased by increasing the output power from the transmitter.
102
Figure 5.40:Measured S11 responses of the OEWG at different temperatures for (a) Si6B1CN
(1000oC) and Si4B1CN (1100
oC).
25oC
1150oC
9.7 9.9 10.1 10.3 10.5 10.79.5 10.9
-50
-45
-40
-35
-30
-25
-20
-55
-15
Frequency (GHz)
S11 a
fter
TD
gating (
dB
)
10.6 10.7 10.8 10.9 11.0 11.1 11.2 11.3 11.410.5 11.5
-35
-30
-25
-20
-15
-40
-10
Frequency (GHz)
S1
1 a
fte
r T
D g
atin
g (
dB
)
25oC
1050oC
(a)
(b)
103
Figure 5.41:Measured resonant frequency and Q of the Si6B1CN (1000oC) sensor
Figure 5.42:Measured resonant frequency and Q of the Si4B1CN (1100oC) sensor
5.4.3 Fabrication and Measurement up to 1300oC
Because the heat pad which was used to control the temperature of the sensor is thin (1.77 cm)
and the heat zone is close to the air, it is hard to achieve the higher temperature and the
25 100 200 300 400 500 600 700 800 900 100010.70
10.75
10.80
10.85
10.90
10.95
11.00
11.05
11.10
f r (G
Hz)
Temperature (oC)
25 100 200 300 400 500 600 700 800 900 100020
30
40
50
60
70
80
90
Q
Temperature (oC)
25 100 200 300 400 500 600 700 800 900 1000 110010.05
10.10
10.15
10.20
10.25
10.30
10.35
10.40
f r (G
Hz)
Temperature (oC)
25 100 200 300 400 500 600 700 800 900 1000 110020
25
30
35
40
45
50
55
Q
Temperature (oC)
104
temperature is limited at 1150oC. Therefore, the furnace which is used for the materials
characterization has to be employed for measuring sensors at higher temperatures.
Figure 5.43:Schematic of the measurement setup for the slot antenna sensor
The schematic of the measurement setup is shown in Fig. 5.43. The furnace is the same as the
one we used for the material characterization. It can rearch the temperature as high as 1500oC
with a high accuracy of ±2oC. The sensor and antenna are in the furnace. The two ends of the
Network Analyzer
Sensor
High Temperature
ConnectorFurnace
Temperature
controller
DC power
supply
Power
cable
Thermocouple
115 Volt
60 Hz Line
Source
High Temperature
Tx / Rx Antenna
105
furnace would be sealed using alumina wool during the measurement. Therefore, the temperature
inside the furnace is pretty stable. The antenna is connected to the network analyzer through a
high temperature cable. The measured results of S11 can be read from the network analyzer.
However, the X-band open-end waveguide can not be directly put into the furnace, because its
structure is bulky. In addition, it can not stand in the high temperature due to its low melting
point metal. So, the high temperature antenna is needed. The high temperature antenna we will
use for the measurement is fabricated on the alumina substrate (0.635 mm in thickness) with the
metal of Pt. The dimension of the designed antenna is shown in Fig.5.44 (a). Fig. 5.44 (b) shows
the simulated S11 of the antenna. The center frequency of the antenna is 10.5 GHz and its
bandwidth is 20.2%, which can cover the range of the temperature dependant resonant frequency
of the sensor.
Figure 5.44:(a) Dimension and (b) simulated S11 result of the antenna. G=23; L1=8.4; L2=8;
W=6.4; h=1.9; s= 0.6; Wf=2.5. (Unit: mm)
G
L
W
h s
L
g
Wf
(b) (a)
106
In order to check the feasibility of this method, a PCB version of the antenna is fabricated and
tested at the room temperature. Rogers 3010 is chosen as the substrate, of which the dielectric
constant (10.2) is close to the one (9.6) of the alumina. The fabricated antenna and the sensor are
shown in Fig. 5.45 (a). The measured result is plotted in Fig. 5.45 (b). It shows the resonant peak
is at 10.53 GHz, which is the same as the previous measurement result. This method works.
Figure 5.45:(a) Fabricated PCB antenna and the high temperature sensor and (b) measurement
result.
The next step is to fabricate the high temperature antenna. First, the outline of the antenna is
printed onto the alumina board and a layer of Pt paste (ESL 5542) is painted to form the antenna
pattern. Then, the alumina board covered by platinum paste are dried at 100ºC for 10 minutes
and sintered at 1000ºC for 10 minutes. The ramp rate during the sintering is 10ºC/min. This
metallization process needs to be repeated at least three times in order to form the enough
thickness of the Pt. At last, Pt layer is approximately more than 30 µm thick. Fig. 5.46 shows the
fabricated antenna. The measured S11 response of this antenna is plotted in Fig.5.46 (b). The
10.0 10.5 11.09.5 11.5
-40
-30
-20
-50
-10
Frquency (GHz)
S11 a
fter
TD
gating (
dB
)
(a) (b)
107
bandwidth is from 9.57GHz to 10.66GHz, which covers the variation range of the temperature
dependant resonant frequency of the sensor.
Figure 5.46: (a) Fabricated high temperature antenna and (b) S11 response of the antenna
Fig. 5.47 illustrates the measurement setup. The high temperature antenna is over the top of the
sensor and used for the wireless sensing, as shown in Fig. 5.47 (a). The distance between them is
kept at 5 mm. One and half inches of the CPW line of the antenna stretches outside the furnace,
so the connector at the end of the antenna feed line is at the low temperature. The inside of
furnace is shown in Fig. 5.47 (b). The sensor, antenna and K-type thermocouple are in the hot
zone of the furnace, where the temperature can be up to 1500oC.
The measurement is performed from 25 to 1300oC. One port Short-Open-Load (SOL) calibration
is done at the end of the coaxial cable, which is connected to the high temperature antenna. The
proper time domain gating is applied onto the measured S parameters. The measured S11 results
of high temperature antenna at different frequency for Si4B1CN (1100oC) sensor are plotted in
9.6 9.8 10.0 10.2 10.4 10.69.4 10.8
-20
-15
-10
-5
-25
0
Frequency (GHz)
S1
1 (
dB
)
(a) (b)
108
Fig.5.48. Each curve represents a S11 response at one temperature in the 25-1300oC range. There
is a unique peak at the highest signal level. This peak corresponds to the resonant frequency of
the sensor. As temperature increases, the resonant frequency and signal level decrease. Above
the temperature of 1300oC, the resonant peak is hard to be observed. According to the
temperature dependant property of the materials, the resonant frequency of the sensor is expected
to decrease monotonously with temperature. However, the resonant frequency at 1350oC is
higher than that at 1300oC, as shown in Fig. 5.49. Because the loss of the Si4B1CN (1100
oC)
material and platinum becomes larger with higher temperature, the re-radiated signal from the
sensor becomes weaker. Therefore, the ripples appear on the S11 curve due to noise. It becomes
harder to detect the resonant frequency at the temperature above 1300oC.
Figure 5.47:Si4B1CN (1100oC) sensor and high temperature antenna (a) outside the furnace and
(b) inside the furnace.
High
Temperature
Antenna
Sensor
High Temperature
Antenna
Sensor Thermocouple
(a) (b)
109
Figure 5.48:Measured S11 responses of the Si4B1CN (1100oC) sensor at different temperatures.
Figure 5.49:Comparison between measured S11 responses of the Si4B1CN (1100oC) sensor at
1300oC and 1350
oC.
9.7 9.8 9.9 10.0 10.1 10.2 10.3 10.4 10.5 10.6 10.79.6 10.8
-40
-30
-20
-50
-10
Frequency (GHz)
S1
1 a
fte
r T
D g
atin
g (
dB
)
1300oC
1350oC
9.7 9.8 9.9 10.0 10.1 10.2 10.3 10.4 10.5 10.6 10.79.6 10.8
-40
-30
-20
-50
-10
Frequency (GHz)
S1
1 a
fte
r T
D g
atin
g (
dB
)
25oC
1300oC
110
The resonant frequencies and Qs of this sensor are extracted and plotted in Fig. 5.50. The
resonant frequency decreases from 10.52 to 9.92 GHz in the temperature range of 25-1300oC,
which corresponds to a measured sensor sensitivity of 0.47 MHz/oC. The highest Q is 34 at room
temperature and the lowest Q is 10, which is observed at 1200oC.
Figure 5.50:Measured resonant (a) frequency and (b) Q of the Si4B1CN (1100oC) sensor
25 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300
0
10
20
30
40
Q
Temperature (oC)
25 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300
9.8
10.0
10.2
10.4
10.6
f r (G
Hz)
Temperature (oC)
(a)
(b)
111
5.5 Conclusion
Based on the temperature-dependant dielectric constant behavior of SiCN materials, three
wireless temperature sensing mechanism have been proposed in this chapter for high-
temperature applications. Corresponding to different antenna configurations, three different types
of sensors are designed and the prototypes are fabricated and tested. All of the sensors
successfully perform at temperatures over 1000oC. These wireless passive sensor designs will
significantly benefit turbine engines in need of sensors operating at harsh environments.
112
CHAPTER 6 SUMMARY AND FUTURE WORK
6.1 Summary
This dissertation focused on the high temperature materials characterization and sensor
applications. Two contributions have been made in this work, which are summarized as follow:
1) A novel method for precisely characterizing the dielectric properties of high temperature
ceramic materials at high temperatures and microwave frequencies are presented in this
dissertation. This technique provides accurate characterization of both dielectric constant
and loss tangent. The dielectric properties of Si(B)CN ceramics are characterized from 25
to 1300oC. Two different metallization processes are implemented for the measurements
with the highest temperatures of 500oC and 1300
oC, respectively. Custom-made through-
reflect-line (TRL) calibration kits are used to maximize the measurement accuracy at
every temperature point. It is observed that the dielectric constant and loss tangent of the
Si(B)CN sample increase monotonously with temperature. This temperature dependency
of Si(B)CN dielectric properties as well as the dielectrically-loaded cavity resonator
structure provides the basis for the development of wireless passive temperature sensors
for high-temperature applications.
2) Wireless temperature sensors are designed based on the cylindrical cavity resonator
loaded with these temperature dependent dielectric materials. Three different types of
sensors are designed and the prototypes are fabricated and tested. All the sensors work
over 1000oC.
113
6.2 Future Work
In order to make the sensors low profile and robust, different materials should be involved as
fewer as possible. This can avoid reaction between different materials and minimize thermal
mismatch between materials, devices and packaging. For the sensors presented in this
dissertation, precious metals are used to cover the ceramic samples to reduce the radiation loss
and therefore form dielectrically loaded high Q resonators. These resonators work at the TM010
cavity mode. For the conventional TE, TM and hybrid dielectric resonator modes, the radiation
loss degrades their Q’s especially for low dielectric constants, such as the one of SiCN, which is
only between 3 and 5. A potentially effective way to achieve high Q and get rid of metals is to
utilize the whispering gallery modes (WGM) of the dielectric resonators. The radiation loss will
decrease when WGM’s are employed, especially with increase of azimuthal mode number [54].
Compared with patch antennas and slots antennas, dielectric image-line antennas would be better
candidates used as receiving/transmitting antennas[55][56], because they are much simpler from
the fabrication’s point of view.
114
LIST OF REFERENCES
[1] M. P. Boyce，Gas Turbine Engineering Handbook, Third Edition, Gulf Professional
Publishing, 2006, Ch. 19.
[2] X. Gong, L. An, and C. Xu, "Wireless passive sensor development for harsh envionment
applications," 2012 IEEE International Workshop on Antenna Technology, Tucson, AZ, Mar. 5-
7, 2012.
[3] G. W. Tregay, P. R. Calabrese, P. L. Kaplin, and M J. Finney, “Optical fiber sensor for
temperature measurement from 600 to 1900oC in gas turbine engines,” Proceedings of SPIE -
The International Society for Optical Engineering, vol. 1589, pp. 38-47, 1991.
[4] T. Tomoki, S. Koji, O. Yoshihiko, and N. Atsushi, “Temperature measurement of high
speed rotating turbine blades using a pyrometer,” Proceedings of the ASME Turbo Expo 2006 -
Power for Land, Sea, and Air, vol. 2, pp. 521-529, 2006.
[5] S. S. Lutz, W. D. Turley, H. M Borella, B. W. Noel, M. R. Cates, and M. R. Probert,
“Remote temperaturemeasurement instrumentation for a heated rotating turbine disk,”
Proceedings of the ISA Aerospace Instrumentation Symposium, vol. 34, pp. 217-229, 1988.
[6] A. Bhattacharya, R. R. Srinivasa, K. M. Ghanashyam, “Characterization of Yb2O3 based
optical temperature sensor for high temperature applications,” Sensors and Actuators, A:
Physical, vol. 134, no. 2, pp. 348-356, 2007.
115
[7] S. Zappe, J. Franklin, E. Obermeier, M. Eickhoff, H. Moeller, G. Kroetz, C. Rougeot, O.
Lefort, and J. Stoemenos, “High temperature 10 bar pressure sensor based on 3C-SiC/SOI for
turbine control applications,”Materials Science Forum, vol. 353-356, pp. 753-756, 2001.
[8] D. J. Young, J. Du, C. A. Zorman, and W. H. Ko, “High-temperature single-crystal 3C-
SiC capacitive pressure sensor,” IEEE Sensors Journal, vol. 4, no. 4, pp. 464-470, August,
2004.
[9] K. Wasa, T. Tohda, Y. Kasahara, and S. Hayakawa, “Highly reliable temperature
sensor using RF-sputtered SiC thin film,” Review of Scientific Instruments, vol. 50, no. 9, pp.
1084-1088, 1979.
[10] P. Nieva, H. Tada, P. Zavracky, G. Adams, I. Miaoulis, and P. Wong, “Mechanical and
thermophysical properties of silicon nitride thin films at high temperatures using in-situ MEMS
temperature sensors,” Materials Research Society Symposium - Proceedings, vol. 546, pp. 97-
102, 1999.
[11] S. V. Patel, M. DiBattista, J. L. Gland, and J. W. Schwank, “Survivability of a silicon-
based microelectronic gasdetector structure for high-temperature flow applications,” Sensors
and Actuators, B: Chemical, vol. B37, no 1-2, pp. 27-35, 1996.
116
[12] R. S. Okojie, A. A. Ned, A. and D. Kurtz, “Operation of α(6H)-SiC pressure sensor at
500°C,” International Conference on Solid-State Sensors and Actuators, vol. 2, pp. 1407-
1409, 1997.
[13] V. V. Varadan, R. D. Hollinger, D. K. Ghodgaonkar, and V. K. Varadan, “Free-space,
broadband measurements of high-temperature, complex dielectric properties at microwave
frequencies,” IEEE Trans. Instrum. Meas., vol. 40, pp. 842-846, 1991.
[14] T. T. Grove, M. F. Masters, R. E. Miers, “Determining Dielectric Constants Using A
Parallel Plate Capacitor,” Am. J. Phys., Vol. 73, No. 1, Jan. 2005
[15] Agilent Application Note, “Agilent Basics of Measuring the Dielectric Properties of
Materials,” Agilent Literature Number 5989-2589en, Jun. 2006
[16] S. Fargeot, D. Guihard, and P. Lahitte, “Dielectric characterization at high temperature
(1600°C) for space applications” IEEE International Conference on Microwave Technology &
Computational Electromagnetics (ICMTCE), 22-25 May 2011, Beijing, China.
[17] S. Bringhurst, M. F. Iskander, and M. J. White, “Thin sample measurements and error
analysis of high temperature coaxial dielectric probes,” IEEE Microwave Theory Tech., vol. 45,
pp. 2073-2083, 1997.
[18] C. A. Jones, “Permeability and Permittivity Measurements Using Stripline Resonator
Cavities: A Comparison,” IEEE Trans. Instrum. Meas., Vol. 48, pp. 843–848, Aug. 1999.
117
[19] X. Gong, W. H. She, E. Hoppenjans, Z. N. Wing, R. Geyer, J. W. Halloran, and W. J.
Chappell, “Tailored and anisotropic dielectric constants through porosity in ceramic components,”
IEEE Trans. MTT, vol. 53, pp. 3638-3647, 2005.
[20] L. Liew, W. Zhang, L. An, S. Shah, R. Lou, Y. Liu, T. Cross, K. Anseth, V. Bright, and
R. Raj, “Ceramic MEMS – new materials, innovative processing and futuristic applications,” Am.
Ceram. Soc. Bull., 80(5), pp. 25-30, 2001.
[21] L. Liew, W. Zhang, V.M. Bright, L. An, M.L. Dunn, and R. Raj, “Fabrication of SiCN
ceramic MEMS using injectable polymer-precursor technique,” Sensors and Actuators: A.
Physical, 89(1-2), pp. 64-70, 2001.
[22] Y. Liu, L. Liew, R. Lou, L. An, V.M. Bright, M.L. Dunn, J.W. Daily and R. Raj.
“Application of microforging in SiCN MEMS structure fabrication,” Sensors and Actuators: A.
Physical, 95 (2-3): pp. 143-151, 2002.
[23] L. Liew, Y. Liu, R. Luo, T. Cross, L, An, V.M. Bright, M.L. Dunn, J.W. Daily, R. Raj.
“Fabrication of SiCN MEMS by photopolymerization of pre-ceramic polymer” Sensors and
Actuators: A. Physical, 95 (2-3): pp. 120-134, 2002.
[24] R. Riedel, G. Passing, H. Schonfelder, and R. J. Brook, “Synthesis of dense silicon-based
ceramics at low-temperatures,” Nature, vol. 35, pp. 714-716, 1992.
[25] R. Riedel, A. Kienzle, W. Dressler, L. Ruwisch, J. Bill, and F. Aldinger, “A silicoboron
carbonotride ceramics stable to 2000 Degrees C,” Nature, vol. 382, pp. 796-798, 1996.
118
[26] L. An, Y. Wang, L. Bharadwaj, Y. Fan, L. Zhang, D. Jiang, Y. Sohn, V. Desai, J. Kapat
and L. Chow, “Silicoaluminum carbonitride with anomalously high resistance to oxidation and
hot corrosion,” Adv. Eng. Mater., vol. 6, no. 5, pp. 337-340, 2004.
[27] Y. Wang, Y. Fan, L. Zhang, W. Zhang, and L. An, “Polymer-derived SiAlCN ceramics
resist to oxidation at 1400oC,” Scripta Mater., vol. 55, pp. 295-297, 2006.
[28] Y. Wang, W. Fei, Y. Fan, L. Zhang, W. Zhang, and L. An, “A silicoaluminum
carbonitride ceramic resist oxidation/corrosion in water vapour,” J. Mater. Res., vol. 21, no. 7,
pp. 1625-1628, 2006.
[29] R. Riedel, L.M. Ruwisch, L. An, and R. Raj, “Amorphous silicoboron carbonitride
ceramics with anomalously high resistance to creep” J. Am. Ceram. Soc., 81, pp. 3341-3344,
1998.
[30] L. An, R. Riedel, C. Konetachny, H.J. Kleebe, and R. Raj. “Newtonian viscosity of
amorphous silicon carbonitride at high temperature” J. Am. Ceram. Soc., 81, pp. 1349-52, 1998.
[31] A. Dhamne, W. Xu, B. Fookes, Y. Fan, L. Zhang, S. Burton, J. Hu, J. Ford, and L. An,
“Polymer-ceramic conversion of liquid polyaluminasilazanes for SiAlCN ceramics” J. Am.
Ceram. Soc., 88[9], pp 2415-19, 2005.
[32] T. Jiang, “Electronic properties and microstructures of amorphous SiCN ceramics derived
from polymer precursors,” Ph.D. Thesis, University of Central Florida, pp. 102-117, 2009.
119
[33] R. G. Nitsche, and E. M. Biebl, “A free-space technique for measuring the complex
permittivity and permeability in the millimeter wave range,” in IEEE MTT-S Int. Microwave
Symp. Dig., pp. 1465–1468, 1994.
[34] V. R. K. Murthy, and R. Raman, “A method for the evaluation of microwave dielectric
and magnetic parameters using rectangular cavity perturbation technique,” Solid State
Communication, vol. 70, no. 8, pp. 847–850, 1989.
[35] A. Kumar, and S. Sharma, “Measurement of dielectric constant and loss factor of the
dielectric material at microwave frequencies,” Progress In Electromagnetics Research, PIER 69,
pp. 47-54, 2007.
[36] X. Ren, T. Jiang, Y. Wang, L. An, and X. Gong, “Characterization of high-temperature
ceramic materials at microwave frequencies for MEMS applications,” 10th IEEE Wireless and
Microwave Technology Conference, Clear Water, FL, April 2009
[37] W. H. She, E. Hoppenjans, and W. J. Chappell, “Embedded dielectric resonator filter in
layered polymer packaging,” IEEE MTT-S Int. Microwave Symp. Dig., pp. 261-264, 2005.
[38] L. Kirkup, and R. B. Frenkel., An Introduction to Uncertainty in Measurement: Using the
GUM (Guide to the Expression of Uncertainty in Measurement), Cambridge: Cambridge
University Press, pp.53-59, 2006.
[39] C. A. Balanis, Advanced Engineering Electromagnetics, New York: J. Wiley & Sons, pp.
494-498, 1989.
120
[40] D. Kajfez, S. Chebolu, M. R. Abdul-Gaffoor, and A. A. Kishk, "Uncertainty analysis of
the transmission-type measurement of Q-factor", IEEE Trans. MTT., vol. 47, pp.367 -371, 1999.
[41] G. E. Ponchak, J. L. Jordan, and M. C. Scardelletti, “High Temperature Characteristics
of Coplanar Waveguide on R-Plane Sapphire and Alumina,” IEEE Trans. Advanced Packaging,
vol. 31, no. 1, pp.146-151, Feb. 2009.
[42] Agilent network analysis applying the 8510 TRL calibration for non-coaxial
measurements, product note 8510-8A, Agilent Technologies.
[43] X. Ren, and X. Gong, “A Wireless Sensing Technique Using Passive Microwave
Resonators,” 2008 IEEE AP-S Int. Symp., San Diego, CA, July, 2008.
[44] H. Chen, “Broadband CPW-fed square slot antennas with a widened tuning stub,” IEEE
Trans. on Antennas and Propagation, vol. 51, no. 8, pp. 1982 – 1986,Aug. 2003.
[45] R. Azadegan, and K. Sarabandi, “Miniature high-Q double-spiral slot-line resonator
filters", IEEE Transactions on Microwave Theory and Techniques, vol. 52, no. 5, pp. 1548 –
1557, May 2004.
[46] X. Ren, S. Ebadi, H. T. Cheng, Y. Chen, L. An, and X. Gong, “Wireless resonant
frequency detection of SiCN ceramic resonator for sensor applications,” 2011 IEEE AP-S Int.
Symp., Spokane, WA, July 2011.
121
[47] X. Ren, S. Ebadi, and X. Gong, “A single-antenna wireless passive temperature sensing
mechanism using a dielectrically-loaded resonator,” in 2012 IEEE AP-S Int. Symp., Chicago, IL,
July 8-14, 2012.
[48] H. Cheng, S. Ebadi, and X. Gong, “A Compact Wireless Passive Temperature Sensor
Using a Resonator/Antenna Integration Up to 1000˚C,” IEEE Antennas and Wireless
Propagation Letter, submitted.
[49] H. T. Cheng, S. Ebadi, X.H. Ren, Y. Yusuf, and X. Gong. “A Compact Wireless Passive
Sensing Mechanism Based on A Seamlessly Integrated Resonator/Antenna,” 2011 IEEE AP-S
Int. Symp., July 3-8, 2011.
[50] H. T. Cheng, Y. Yusuf, and X. Gong, “Vertically Integrated Three-Pole Filter/Antennas
for Array Applications,” IEEE Antennas Wireless Propagat. Lett., vol.10, pp.278-281, 2011.
[51] Y. Yusuf, and X. Gong, “Compact low-loss integration of high-Q 3-D filters with highly
efficient antennas,” IEEE Trans. on Microwave Theory Tech., vol. 59, no. 4, pp. 857-865, Apr.
2011.
[52] Y. Yusuf, H. T. Cheng, and X. Gong, “A seamless integration of 3-D vertical filters with
highly efficient slot antennas,” IEEE Trans. Antennas Propagat., vol. 59, no.11, Nov. 2011.
[53] H. A. Haus, Waves and fields in optoelectronics, Prentice-Hall, 1984, ch.7.
122
[54] M. Neshat, D. Saeedkia, S. Gigoyan, and S. Safavi-Naeini, “Mode selective dielectric
resonator coupled to dielectric image waveguide for sensing applications,” in Proc. 32nd Int.
Conf. Infrared Millimeter Waves, Cardiff–Wales, U.K., pp. 764–765, Sep. 2007.
[55] Song Lin, Joshua L. Wilson, Aly E. Fathy and Suren Gigoyan, “Development of
millimeter wave antennas and associated feed network components using dielectric image guide,”
International Journal of RF and Microwave Computer-Aided Engineering, vol. 20, pp. 56–65,
Jan. 2010.
[56] S. Gigoyan, D. Saeedkia, M. Neshat, H. Chen, and S. Safavi-Naeini, “Tapered Dielectric
Image-Line Antenna Array for Millimeter-Wave Applications,” 2008 IEEE Radio and Wireless
Symposium , pp.667-670, Jan. 2008.